LSMS GUIDEBOOK September 2021 Agricultural Survey Design Lessons from the LSMS-ISA and Beyond Andrew Dillon, Gero Carletto, Sydney Gourlay, Philip Wollburg and Alberto Zezza LSMS GUIDEBOOK September 2021 Agricultural Survey Design Lessons from the LSMS-ISA and Beyond Andrew Dillon, Gero Carletto, Sydney Gourlay, Philip Wollburg and Alberto Zezza ABOUT LSMS The Living Standards Measurement Study (LSMS), a survey program housed within the World Bank’s Develop- ment Data Group, provides technical assistance to national statistical offices in the design and implementation of multi-topic household surveys. Since its inception in the early 1980s, the LSMS program has worked with dozens of statistical offices around the world, generating high-quality data, developing innovative technologies and improved survey methodologies, and building technical capacity. The LSMS team also provides technical support across the World Bank in the design and implementation of household surveys and in the measure- ment and monitoring of poverty. ABOUT THIS SERIES The LSMS Guidebook series offers information on best practices related to survey design and implemen- tation. While the guidebooks differ in scope, length, and style, they share a common objective: to provide statistical agencies, researchers, and practitioners with rigorous yet practical guidance on a range of issues related to designing and fielding high-quality household surveys. The series aims to achieve this goal by draw- ing on the experience accumulated from decades of LSMS survey implementation, the expertise of LSMS staff and other survey experts, and new research using LSMS data and methodological validation studies. Copyright © 2021 The World Bank. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http://creativecommons.org/licenses/ by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following condition: Attribution Please cite the work as follows: Dillon, A., Carletto, G., Gourlay, S., Wollburg, P., & Zezza, A. (2021). Agricultural Survey Design: Lessons from the LSMS-ISA and Beyond. Washington DC: World Bank. Disclaimer The findings, interpretations, and conclusions expressed in this Guidebook are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Living Standards Measurement Study (LSMS) World Bank Development Data Group (DECDG) lsms@worldbank.org https://www.worldbank.org/lsms data.worldbank.org Cover images (from top to bottom): © Ray Witlin/World Bank, © Tomas Sennett/World Bank, © Central Statistical Agency of Ethiopia, © Aisha Faquir/World Bank AGRICULTURAL SURVEY DESIGN iii TABLE OF CONTENTS 1. INTRODUCTION...............................................................................................................................................1 2. SURVEY METHODOLOGY FOR AGRICULTURAL DATA COLLECTION...................................................3 Innovations in Research on Survey Design Choices........................................................................................................................ 4 Survey Design Choices: Units of analysis for agricultural measurement.....................................................................................7 Survey Design Choices: Respondents.................................................................................................................................................9 Survey Design Choices: Recall periods.............................................................................................................................................10 Survey Design Choices: Minimizing attrition in longitudinal surveys.........................................................................................11 3. PRODUCTION..................................................................................................................................................12 Land Measurement................................................................................................................................................................................15 Field Crop Production and Yield Measurement: Self-reported, remote-sensing, and crop-cut production measures....19 Agricultural Production Data: Design Features of the Reference Questionnaire...................................................................23 4. AGRICULTURAL INPUTS................................................................................................................................24 Agricultural labor...................................................................................................................................................................................24 Seed, fertilizer, pesticides and herbicides .........................................................................................................................................25 Agricultural Inputs: Design Features of the Reference Questionnaires....................................................................................28 5. LIVESTOCK.......................................................................................................................................................30 Measuring Stocks...................................................................................................................................................................................30 Costs of Production .............................................................................................................................................................................31 Income sources......................................................................................................................................................................................31 Challenges to Livestock Measurement.............................................................................................................................................32 Livestock Data: Design Features of the Reference Questionnaires...........................................................................................34 6. FIELD IMPLEMENTATION...............................................................................................................................35 Fieldwork Implementation Tradeoffs and Survey Design Choices..............................................................................................35 Piloting for Improved Survey Design.................................................................................................................................................37 Mode of Data Collection.....................................................................................................................................................................37 Fieldwork Organization and Logistics...............................................................................................................................................38 Plot Visits and Georeferencing............................................................................................................................................................39 7. CONCLUSIONS AND FUTURE DIRECTIONS FOR AGRICULTURAL SURVEY DESIGN.......................41 Measuring Theoretical Concepts More Precisely...........................................................................................................................42 Open Measurement Questions for Validation Research .............................................................................................................43 REFERENCES.........................................................................................................................................................44 iv AGRICULTURAL SURVEY DESIGN APPENDIX I. GLOSSARY.............................................................................................................................................................51 APPENDIX II. AGRICULTURAL REFERENCE QUESTIONNAIRE...........................................................................................55 APPENDIX III. LIVESTOCK REFERENCE QUESTIONNAIRE..................................................................................................119 List of Tables TABLE 1. SURVEY DESIGN CHOICES............................................................................................................................................... 5 TABLE 2. SURVEY DECISIONS THAT INFORM PRODUCTION MEASUREMENT.............................................................13 TABLE 3. RESEARCH ON LAND AREA MEASUREMENT.........................................................................................................18 TABLE 4. RESEARCH ON PRODUCTION MEASUREMENT....................................................................................................22 TABLE 5. SURVEY DECISIONS THAT INFORM INPUT MEASUREMENT.............................................................................27 TABLE 6. DESIGN DECISIONS IN LIVESTOCK PRODUCTION MEASUREMENT............................................................33 AGRICULTURAL SURVEY DESIGN v ACKNOWLEDGMENTS The authors are grateful to Didier Yelognisse Alia, Leigh Anderson, and Paul Winters for their valuable peer review of this document, as well as the continued efforts and dedication of the NSOs involved in the Living Standards Measurement Study – Integrated Surveys on Agriculture program. Much of the research and field- work underpinning this document was made possible thanks to the generous financial contributions of the Bill and Melinda Gates Foundation, the United Kingdom’s Foreign, Commonwealth & Development Office (formerly FDID), and the United States Agency for International Development. vi AGRICULTURAL SURVEY DESIGN ACRONYMS AND ABBREVIATIONS APIs Application programming interfaces CAPI Computer assisted personal interviewing CR Compass and rope CSPro The Census and Survey Processing System GPS Global Positioning System ICLS International Conference of Labour Statisticians LSMS-ISA Living Standards Measurement Study - Integrated Surveys on Agriculture NGO Non-governmental organization NSO National Statistical Office ODK Open Data Kit PAPI Paper assisted personal interviewing 1 1. Introduction Household surveys are the primary tool through be measured, along with determinants of poverty to which international development goals are monitored, facilitate policy analysis by NSOs and relevant national and policy questions are informed based on empirical ministries and policymakers. Concurrently, an academic research. Accurate agricultural data are predicated on interest in understanding household consumption in careful survey design and data collection, though practices low- and middle-income countries led to research foun- for survey design and implementation of data collection dations for measuring household consumption aggregates exercises vary considerably across national statistical (Deaton and Zaidi, 2002) and the canonization of the offices (NSOs), international organizations that generate agricultural household model (Singh et al., 1986) as the international data public goods, NGOs, and researchers. relevant unit of observation for understanding small- While the importance of agricultural statistics is widely holder production and consumption decisions. These acknowledged as essential to evidence-based policy- insights provided a basis for survey design for multi-topic making, tracking progress in the agricultural sector, and household surveys focusing on linkages between produc- conducting innovative research, best practices in data tivity changes (human capital, labor productivity and later collection and the research foundations of these rec- agricultural productivity) and welfare which are essential ommendations are often ambiguous. Data collection to policy analysis. A focus on household production also methods may not be completely generalizable and need facilitates supply and demand comparisons for food secu- to be adapted to the local context, but the lessons rity and agriculture sector analysis. learned across contexts and the lessons that we can gen- eralize, can be available to the community of researchers Out of the lessons of the past 18 years since the Grosh and practitioners who collect and use agricultural data. and Glewwe volumes, innovations in methods and mea- It is the purpose of this volume to aggregate innovations surement have improved our understanding of best in agricultural survey design from the experience of the practices in multi-topic household survey design. A widely Living Standards Measurement Study - Integrated Surveys noted limitation of the previous volumes and multi-topic on Agriculture (LSMS-ISA), academia, and other survey household surveys generally has been agricultural module operations to provide recommendations for research- recommendations. Though acknowledged as a focal liveli- ers, survey practitioners and policy analysts on lessons hood practice of the world’s poor, multi-topic household learned in survey design, while also highlighting topics for surveys designed to measure poverty have often been which more research is required to provide a conclusive limited by cursory agricultural data in part due to the survey design recommendation. From these agricultural burden of measuring consumption aggregates. survey design ‘lessons learned’, we highlight tradeoffs between household survey design choices, measurement Rozelle (1991) described three differing approaches to error and data use. agricultural survey design including a production function approach, an income statement approach and a balance In their synthesis of household survey practices, Grosh sheet approach. A production function approach places and Glewwe (2000) outlined lessons learned from 15 emphasis on capturing how inputs create outputs for an years of the World Bank’s Living Standards Measurement agricultural household unit with an aim to potentially esti- Study (LSMS) surveys. The genesis of this investment in mate the returns to different inputs on farm productivity. the design and implementation of household surveys An income statement approach measures farm revenues was motivated by the realization that progress towards and expenses with an objective of measuring farm profits.A reducing poverty could only be achieved if the goal could balance sheet approach is related to the income statement 2 AGRICULTURAL SURVEY DESIGN approach but differs in valuing farm assets and liabilities as In Chapter 3, we discuss the design of agricultural produc- well as the sources of all inputs and outputs. Agricultural tion modules, particularly as they relate to the organization, sector specific surveys often take alternative approaches unit of analysis, and differences in production modules by to using the farm or area sampling as the unit of analysis agricultural system (field crops, root crops, and agro-for- rather than the household (Benedetti et al., 2010). estry). A key innovation in the design of agricultural surveys, which is discussed in Chapter 3, is the enumera- A recent innovation in survey design has been the LSMS-In- tion of land holdings of the household by agricultural plot tegrated Surveys on Agriculture which have redesigned and plot manager. This survey design choice improves the agricultural modules by recognizing first, that the unit of measurement and analysis of agricultural yields and gener- observation for agricultural production is often not the ates sex-disaggregated agricultural data but creates survey household, but the plot that may be managed by differing design tradeoffs in the respondent’s ability to recall alloca- household members, and second, that improved agricul- tion of inputs to plots which may be purchased in bulk to tural statistics may be collected and respondent survey distribute across plots (fertilizer) or be difficult to retro- burden reduced by timing agricultural data collection to spectively recall (household labor hours or days per plot). the agricultural production seasons in a post-planting and post-harvest questionnaire. Multiple time periods facilitate Chapter 4 addresses the design of modules to measure agricultural input and production recall, while also low- agricultural factors of production given plot-level report- ering the per interview time for a multi-topic household ing including labor, capital, inputs (fertilizer and seed), and questionnaire. Repeated survey visits open the possibility water management. Chapter 5 covers livestock production of different panel types where the unit of analysis could for which previous multi-topic survey design attention be either the household, plot or farmer. The LSMS-ISA has been limited. As low- and middle-income countries’ surveys track households and plot managers over time, economies have grown, animal-based sources of food pro- but not plots due to high plot visit costs and challenges duction independent of cereal and tree crop production related to plot recall over time. are increasingly important. Motivation for measuring live- stock or animal holdings in previous multi-topic surveys Improved agricultural data opens a wide range of policy has often focused on livestock as an asset stock, yet pasto- analysis for both research and policy. NSOs are common ralists rely on animal rearing not only as a stock of wealth, implementers of agricultural surveys and require high but also as a flow of income. Measuring revenues and costs quality agricultural data to meet a number of national from animal production has been challenging given the policy objectives, including annual or seasonal estimation frequent and often small costs associated with animal rear- of crop production and monitoring change in agricultural ing and the imperfectly measured effort and time that are production over time. Agricultural surveys, if carefully often important determinants of animal health and quality. designed, can meet these objectives while also enabling additional research on the drivers of agricultural pro- In Chapter 6, we provide practical guidance for the col- duction and productivity, for example, which can inform lection of agricultural data in a multi-topic household policies aimed at improving agricultural production and survey including field work logistics, some perspectives rural livelihoods. of implementers including NSOs, NGOs, and research- ers. To conclude, we summarize key lessons learned and This volume is organized to assess innovations in agricul- highlight future methodological research which may be tural survey design, provide practical recommendations promising to advance agricultural survey design meth- on the design and tradeoffs inherent in agricultural survey odology in Chapter 7. In the Appendices and online, modules, and lastly to provide lessons learned on the we provide a reference questionnaire, which we use as organization of fieldwork and the linkage of agricultural a basis for discussing agricultural survey design choices household data to other data sources. In the next chapter, throughout this volume1. Chapter 2, we assess recent methodological research con- ducted by the LSMS team and the academic community to inform current best practice recommendations.We discuss topics including questionnaire design, respondent selection, units of analysis, timing of data collection to understand 1 Editable versions of the Agricultural Survey Design reference seasonality, recall periods, and innovations in the measure- questionnaires are available for download at: https://www.worldbank. ment of production, land, and agricultural labor. org/en/programs/lsms/publication/AgriculturalSurveyDesign 3 2. Survey Methodology for Agricultural Data Collection This chapter reviews the general insights from the survey survey response, a researcher faces additional decisions design literature and recent innovations in agricultural and tradeoffs in the research design process. Minimizing data collection which inform agricultural module design total survey error with expensive measurement methods choices. It is important to remember that survey design ignores the research design cost-variance tradeoff and the for multi-topic household surveys is a relatively new and full set of research design choice variables. For example, interdisciplinary endeavor. De Heer et al. (1999) discuss in the case where the researcher completely conceives the history of survey design, documenting how innova- of the research design, a researcher chooses an identifi- tions in sampling methods, data collection techniques, and cation strategy along with the questionnaire design and statistical methods for data analysis were developed in a sample size using a sampling strategy given their budget the 1930s. The early field of survey research emphasized constraint. A researcher may be willing to accept some limiting measurement error and nonresponse through measurement error if reducing such error reduces the protocols developed in surveys conducted by Bowley statistical power of the research design. If a researcher is and Burnett-Hurst (1926) to measure poverty among the implementing a randomized controlled trial, measurement working class in England. Agricultural measurement and error that is not correlated with treatment status may related issues of non-random measurement error have not bias estimates, whereas in a non-experimental design been discussed in some of the earliest work by Fisher measurement error might bias parameter estimates. (1926) and Working (1925),2 but also have important overlap with survey methodology as discussed in the To conceptualize more carefully these research design statistics, applied econometrics, development economics, tradeoffs and to emphasize the point that question- and labor economics fields. naire design is one set of choices among many in the research design, which include an identification strategy, The survey methodology literature has focused on sampling strategy and field implementation protocol, minimizing bias relative to the cost of improved data Dillon et al. (2020) introduce the idea of the data qual- collection in framing the questionnaire design problem ity production function. The researcher’s objective is to as minimizing total survey error (Gideon, 2012). Total maximize knowledge or evidence generated from the survey error can be disaggregated into sampling error research project. The researcher makes choices about and non-sampling error, where this volume focuses pri- the identification strategy, statistical power and external marily on the latter. The questionnaire design affects validity that are subject to budget constraints and the non-sampling error, which can be further disaggregated data quality production function. The data quality pro- into non-response errors (for example refusals, miss- duction function includes questionnaire design choices, ing data, respondent not knowing) and response errors but also interacts with the other choice variables such (such as question framing, question sequencing, and social as sampling, empirical approach and field implementation desirability bias). While this framework is useful in con- protocols and constraints. From the perspective of NSOs sidering sources of sampling and measurement error in designing surveys for both policy and potential research purposes, additional constraints need to be considered when designing agricultural surveys, including financial 2 For a deeper discussion of the evolution of methods and measurement resource availability, personnel capacity, and competing in agricultural economics, see Fox (1986) and Herberich et al. (2009). 4 AGRICULTURAL SURVEY DESIGN demands on and mandates of the agency. Considerations generate as much knowledge as possible. We then focus for designing surveys to meet the needs and constraints on questionnaire design choices that researchers make of NSOs is further discussed in Chapter 6. including the units of analysis of agricultural measurement, respondent choice, recall periods, module sequencing, The history of the Living Standard Measurement Study question phrasing, and panel data design. We outline the surveys began in the 1980s as part of a concerted effort theoretical motivation for some of the design choices and at the World Bank to provide global poverty measures how innovations in questionnaire design have addressed and support the statistical capacity of developing coun- non-classical measurement error. In our review of specific try governments to collect household data. The Grosh modules in subsequent chapters, we emphasize question- and Glewwe (2000) manuals on questionnaire design naire design choices that may reduce measurement error, were an important consolidation of field knowledge and but also highlight where choices represent tradeoffs in research on questionnaire design that resulted from the research design whose objective may not be to solely Living Standard Measurement Study survey program. An reduce measurement error if identifying a specific empiri- early innovation in the LSMS approach and in papers cal objective is compromised due to cost tradeoffs. such as Deaton and Zaidi (2002) was the clear discus- sion between the theoretical motivation for estimating INNOVATIONS IN RESEARCH ON an empirical relationship and its linkage to questionnaire SURVEY DESIGN CHOICES design in a multi-topic household survey. Since these semi- nal works, research on questionnaire design has continued Given the multitude of potential questionnaire design to improve the internal validity of empirical estimates choices, how does a researcher or NSO decide which through improved measurement and data quality that choices are most important? While field testing and pilot- reduces bias from non-random measurement error. ing strategies are often used to improve survey design, an emerging literature uses survey experiments or vali- In the late 2000s, earlier surveys with limited agricul- dation studies to inform either the relative or absolute tural modules used household level reporting but did biases of survey design choices. De Weerdt et al. (2019) not establish a clear linkage between agricultural inputs provide a recent summary of the survey experiment and and outputs. The inherent tradeoff in a multi-topic validation literature. Survey experiments randomly assign household survey is how to maximize potential empir- survey design choices to assess their relative causal effect. ical analysis within and across modules given survey Random assignment provides strong internal validity which budget and respondent fatigue constraints. Carletto et al. is important due to multiple design choices contained (2010, 2015) outlined the methodological innovations in in a multi-topic questionnaire. However, the strength of designing integrated agricultural modules in a multi-topic this approach is also a potential limitation. In particular, household survey. If surveys were going to inform food the measurement error associated with each choice is price policy, poverty reduction in rural areas, or agricul- not measured against ‘objective truth’. We can know the tural policy to improve farm productivity, agricultural tradeoffs between design choices and their magnitude, but modules needed to better measure the relationship both design choices could be measured with error. between inputs and outputs at the plot level, measure changes in agricultural livelihoods over time and account Validation studies rely on comparing alternative designs to for seasonality. These innovations raised a host of survey a predefined truth such as administrative data, an objec- design considerations and the potential to integrate tive measure or repeated observations over time. The cost-effective technologies through computer assisted advantage of this survey methodology ‘methodology’ is personal interviewing (CAPI) and GPS devices both of that the causal effect of survey design choice is estimated which necessitated a research program using survey relative to this benchmark, quantifying the measurement design experiments to better inform the tradeoffs and error associated with each choice. The limitation is that potential biases from survey design choices. Below, we ‘objective truth’ might not be measured in administra- review what has been learned from these survey exper- tive data if it too is reported with error or the objective iments and the broader literature since the Grosh and measure itself may not have been validated. When vali- Glewwe (2000) agricultural chapter. dation studies rely on repeated observations to assess consistency of responses, assuming that convergence We first present a framework for thinking about research reflects the true survey response, the effect of repeated design and questionnaire design where the objective is to interviewing may bias estimates. In the remainder of this 2. Survey Methodology for Agricultural Data Collection 5 document, we focus on methodological innovations that in the survey data. If estimates of item nonresponse and have been validated either through survey design exper- response errors could be known from a pilot, a question- iments or validation studies with an objective measure naire design could make a comparison between the bias that has either been deemed the gold-standard approach induced by skip sequences and the survey time cost. or validated against the relevant gold-standard. Methodological research has focused on question phras- A survey design literature outlines the importance of ing including ordering screening questions to make module sequencing, question ordering and question ambiguous concepts clearer or using different word phrasing (Iarossi, 2006) in household survey design choices. An earlier survey design literature established that is also relevant, but not specific to, agricultural the importance of question phrasing (Payne 1980; Sudman survey design. Surveys should follow a logical sequence, and Bradburn 1973; Converse and Presser 1986; Fowler collect relevant information from the best-informed 1995; Gideon 2012). General recommendations from this respondent, and be coherent to the respondent. Little literature focused on the way respondents understand methodological research has been implemented on questions in a face-to-face interview. Misunderstanding module sequencing as it is presumed feedback from questions from the respondent’s perspective can result piloting that detailing individual level outcomes within from ambiguity, framing, technical wording, or hypothet- the household via the household roster, education and ical construction. Bautista (2012) describes a theoretical labor modules, for example, should precede household model of the survey response process in four steps: the level outcomes such as consumption, non-food con- respondent’s comprehension of the question, the cog- sumption, and household enterprises. nitive process of information retrieval from memory, judgement of the appropriate answer, and communica- Manski and Molinari (2008) framed the choice of skip tion of the answer. A respondent’s answer to questions sequencing in a module as a formal design process. When can also be misreported by enumerators if response including skip sequences, a questionnaire design makes a categories or the question’s intention is not clear to the survey design choice that may increase item nonresponse enumerator, which may result in incorrect administration and response errors leading to non-random measure- of the questionnaire. ment error. Manski and Molinari (2008) illustrate that the tradeoff in using skip sequences which reduce respon- We summarize key agricultural survey design choices dent burden depends on the likelihood and the cost of in Table 1, describing these design choices in detailed item nonresponse and response errors that increase bias subsections. Table 1. Survey Design Choices SURVEY DESIGN CONSIDERATIONS KEY REFERENCES CHOICE KNOWLEDGE GAPS Sampling unit: Household: representative estimate of household Integration of farm Carletto et al., 2010; household vs welfare, household-run farms, but not of entire and household Global Strategy to agricultural holdinga agriculture sector; may also not be the right choice for frames improve Agricultural and transhumant pastoralists. Rural Statistics, 2017; Holding: representative estimates of agriculture sector; FAO 2015a; FAO 2015b. no direct link to household welfare; not necessarily the best choice for livestock, forestry, aquaculture production. A further consideration is the availability of recent sampling frames, that is, lists of households or holdings. Household frames are usually based on the most recent population and housing census and tend to be more widely available than up-to-date list frames of agricultural holdings which may be based on agricultural registers or the most recent agricultural census. 6 AGRICULTURAL SURVEY DESIGN SURVEY DESIGN CONSIDERATIONS KEY REFERENCES CHOICE KNOWLEDGE GAPS Definitions of Consumption: individuals must regularly share meals to Impact of Beaman and Dillon, the household: be members of the same household. household 2012; production-based vs Production: Adult individuals co-mingle revenues from definition Guirkinger and consumption-based; agricultural production and non-farm enterprise for agricultural land Platteau (2014); wording of definition consumption. holding and production Kazianga and Wahhaj Issues such as (seasonal) migration and polygamy (2013) may complicate the concept of household and varying definitions may affect the measured size and composition of households and have an impact on recorded agricultural production, among others. To produce comparable and reliable agricultural data, surveys need to use comparable and suitable household definitions. Unit of observation Plot/parcelb: production technologies, ownership, and Quantified impact Just and Pope, 2001; for crop production productivity differ across plots within a household, so of measuring at plot Carletto et al., 2017; measurement: that the plot is a more appropriate unit of observation vs household level Plot/parcel vs to match outputs with inputs, and for understanding on reported inputs Carletto et al., 2010 household/holding dynamics within the household; labor and fertilizer and outputs, and input allocation at plot-level may be more burdensome how they compare to recall, while land input, output, and tenure is likely to an objective measured more reliably at plot-level. measure Plots may change season-on-season, so that tracking plots over time is difficult, while parcels are expected to remain constant. Household/holding: easier to recall when inputs are purchased collectively rather than separately for each plot/parcel; potentially reduces questionnaire length; Respondent Self/individual: individual reporting considered more Impact of proxy- Ambler et al., 2017; selection: reliable than proxy reporting, especially when it comes reporting Bardasi et al., 2011; self vs proxy to capturing gender dynamics; interviewing each Coates et al., 2010; respondents; group respondent individually allows for truthful reporting; interviews vs may lead to inconsistent reports between different Dillon et al., 2012; individual interviews respondents; Dillon et al., 2021; Proxy reporting: less time intensive and Doss et al., 2019; methodologically challenging; may lead to overlooking Jacobs and Kes, 2015; individuals, their contributions and vulnerabilities Janzen, 2018; Kilic et al., 2020b; Kilic and Moylan, 2016; Palacios-Lopez et al., 2017; Serneels et al., 2017 Recall periods: Length of recall period: important to match recall Effect of Beegle et al., 2012; length vs number of period with activity appropriate time horizon; longer repeated visits on Gaddis et al., 2019; field visits recall periods increases cognitive burden of recalling responses (such as through learning, Kilic et al., 2018; events, especially when these are less salient; very short recall periods may lead to double counting; experience) Wollburg et al., 2020  Number of field visits: more visits imply shorter recall periods for outputs and inputs, as do shorter implementation periods; more visits increase fieldwork costs. 2. Survey Methodology for Agricultural Data Collection 7 SURVEY DESIGN CONSIDERATIONS KEY REFERENCES CHOICE KNOWLEDGE GAPS Longitudinal surveys: Choice of panel unit: panel may be at the level of Effects of long-term Garlick et al., 2020; Panel unit; attrition, the enumeration area, household/farm, parcel, plot, panels on sample of Bevis and Barrett, 2020; representativeness, each allowing for more sophisticated analysis but also farmers and their characteristics Schündeln, 2018; and tracking; increasing logistical difficulty and potentially costs; conditioning plot is expected to change season-on-season and may Zwane et al., 2011 therefore not be a suitable panel unit. With any choice of panel, attrition is a concern as it leads to attrition bias; tracking split-off households is advisable but increases costs and logistical challenges. A representative sample is representative at the time it is drawn, but changes to populations can lead panel samples to lose representativeness over time, which may be aided through refreshing the sample. There are also potential response effects from re-visiting the same households, panel conditioning. a According to the FAO World Program of the Census of Agriculture (WCA), the agricultural holding is defined as an “economic unit of agricultural production under single management comprising all livestock kept and all land used wholly or partly for agricultural production purposes, without regard to title, legal form or size.” b A parcel is any piece of land of one land tenure type entirely surrounded by other land, water, road, forest or other features not forming part of the holding, or forming part of the holding under a different land tenure type. A plot is defined as a continuous piece of land on which a specific crop or a mixture of crops is grown or which is fallow or waiting to be planted. SURVEY DESIGN CHOICES: poverty, and other facets of rural livelihoods. As a nation- UNITS OF ANALYSIS FOR ally representative household survey, the household as AGRICULTURAL MEASUREMENT the unit of analysis is inherently microeconometric in approach as opposed to providing a sector-representa- The unit of analysis is critical for the validity and compa- tive sample. By using the household as the unit of analysis, rability of empirical estimates. In multi-topic household we do not estimate agricultural sector indicators, as the surveys that include welfare measures based on house- sample will exclude farm enterprises that operate in the hold consumption as well as production, several issues private sector as opposed to within agricultural house- related to the units of analysis must be addressed in the holds (Table 1). context of a nationally representative survey. How exactly is the household defined in the context of Households extended and nuclear family overlapping responsibilities in production, income sharing, and social connected- In defining the population frame, is the relevant unit of ness? Household definitions differ widely across surveys analysis the household, farm, holding, or enterprise? and have important measurement implications in part A national survey of the agricultural, manufacturing or because social and economic definitions of the house- industrial sector would not necessarily focus on the hold diverge (Beaman and Dillon, 2012), especially in household as the relevant unit of analysis. Agricultural communities with extended farming families with land censuses use a population frame defined by agricultural inheritance claims, common production of family lands, holdings, acknowledging that both small-scale farms oper- or complicated use-rights. Key words in the definition ated by the majority of rural farmers and agri-industrial of the household in multi-topic household surveys vary farms of larger scale contribute to the agricultural sector. depending on their survey purpose and can emphasize Using a household population frame would inherently either a family structure such as the nuclear family, con- exclude larger scale enterprises and farms that are not sumption requirements such as eating from a ‘common operated by households. As the analytical objective of pot’, or production based definitions that include the LSMS surveys has focused on poverty measurement requirements that adults co-mingle revenues from agri- and the determinants of changes in household welfare, cultural production and non-farm enterprise for mutual the household is the relevant unit of analysis which can consumption. The household definition matters from still inform how policy changes agricultural productivity, an agricultural perspective as who is included in the 8 AGRICULTURAL SURVEY DESIGN household will be included in modules on agricultural Tradeoffs in measurement error related to different land holdings, labor, assets, and marketing decisions. In units of analysis also imply different empirical applica- polygamous households, multiple adult farmers, compet- tions. Nationally representative surveys such as FAO’s ing land use rights, and collective agriculture for some Agricultural Census provide useful agricultural informa- plots may be key features of the agricultural production tion which may be preferred to a household or plot level system that nuclear definitions of the household may analysis if the researcher wants to make sector specific exclude. Empirical studies of polygamy and agriculture analysis related to farm size when a country has large have incorporated wider definitions of the household scale commercial farms, infer changes in agricultural labor to measure the complexities introduced in production markets, or analyze changes in agricultural businesses. (Guirkinger and Platteau, 2014; Kazianga and Wahhaj, LSMS survey design is motivated by empirical analysis 2013). Definitions of the household that exclude agricul- that can be tested using an agricultural household model tural producers lower reported agricultural production. (Singh et al., 1986). Residency requirements also complicate measurement of pastoralist activities when transhumant pastoralism is Across different agricultural systems, the vocabulary used by the household. associated with an agricultural landholding may also differ. Farmers use different words to indicate their farm, Beaman and Dillon (2012) show that consumption and parcel and plot, often with contradictory meanings. It is production key words in the definition of the house- important that agricultural survey design reflect a clear hold do not alter household size estimates, but do affect conception of the hierarchy of units consistent with the household composition. Hess et al. (2001) describe the agricultural system that is being measured. Carletto et al. results of the Census Bureau’s Questionnaire Design (2016) provide an overview of land area measurement Experimental Research Survey where individual- and survey design issues, noting differences in units of land household-level measures were collected. They find that measurement as well as variation across LSMS-ISA sur- with the exception of asset data and health insurance veys in land reporting units. The holding, parcel, field and prevalence reporting that household level enumeration plot have internationally accepted definitions, though of survey questions provided equivalent estimates to that their interpretations by both academics, NSOs, and policy of individual level data for demographic and government makers often lead to ambiguity. transfer reporting. Given the costs of collecting individual level data, they conclude that household level data are The definition of the agricultural holding is the primary preferable, but the type of variable likely affects data qual- unit of analysis in agricultural surveys, whereas the house- ity and differences in reporting biases attributable to the hold is the primary unit of analysis in household surveys. unit of analysis (Table 1). FAO (2015a) defines the agricultural holding as,“economic unit of agricultural production under single management Plots comprising all livestock kept and all land used wholly or partly for agricultural production purposes, without regard In the design of agricultural modules, household level to title, legal form or size...” Holdings can be divided into measurement of agricultural inputs and outputs did parcels and the FAO notes that the difference between not correspond well with production technologies and the holdings land sub-units are distinguished such that “A reasonable assumptions about these technologies as field is a piece of land in a parcel separated from the rest represented in the literature on production and profit of the parcel by easily recognizable demarcation lines.... A function estimation (Chambers 1988; Chambers and field may consist of one or more plots, where a plot is a Quiggin 2000; Just and Pope 2001). An important survey part or whole of a field on which a specific crop or crop design choice has been selecting a better unit of analysis mixture is cultivated, or which is fallow or waiting to be than the household to measure agricultural production. planted.” (FAO, 2015a). When designing and implementing Measuring productivity and input use at the plot is consis- an agricultural survey, practitioners should confirm the tent with the production literature, but multiple cropping, tiers and definitions used by the national statistical office seasonality, and distinctions between plot owners and as they may not coincide with the FAO definitions. Note plot managers create new measurement challenges in that the reference questionnaire included in Appendix 2 attributing inputs to agricultural outputs. utilizes only the parcel and plot tiers, as this is common in many contexts. 2. Survey Methodology for Agricultural Data Collection 9 Recording agricultural information at the household SURVEY DESIGN CHOICES: level inherently aggregates individual production and RESPONDENTS imposes a linearity assumption across plots for input utilization and asset use. The main tradeoff in recording A larger discussion within the survey design literature is agricultural information at the plot level is that farm- estimating the potential bias from self-reporting versus ers must recall input allocation at the plot level which proxy reporting (Moore, 1988; Krosnick, 1999). In labor requires more cognitive effort and response time. These modules that collect earnings data, it is presumed that recall biases may be compounded by proxy response self-reported data may be the most accurate as respon- bias as plot level self-reporting is time consuming in the dents have full information regarding their earnings which field and may be not feasible for all survey responses. may not be fully disclosed to other household members. Proxy respondents may have incomplete information on This assumes that self-respondents do not have incen- plots managed by other household members. For farm- tives to mis-report information if they think that lower ers who purchase inputs collectively with their family earnings increase their probability of future program for multiple plots, it may be difficult to accurately assess participation or transfer. Bardasi et al. (2011) found in how much fertilizer, seed, or other input was applied the context of labor market modules that using proxy to a particular plot. Consistent with time use data, it responses as opposed to individual self-reports had no may also be difficult for a farmer to recall individual effect on female labor statistics, but that proxy reports household labor allocations to particular plots over did lead to under-reporting of men’s participation rates an agricultural season or with respect to particular in agricultural activities. Kilic et al. (2020a) also find signif- agricultural tasks. While more research is needed to icant impacts of respondent strategy in the collection of understand the measurement implications of the disag- labor data in their analysis of two concurrent surveys in gregation of input and production data to the plot level Malawi. They find that the use of the common approach from the household level, the known analytical advan- which allows proxy and non-private interviewing results tages of doing so, such as analysis of male-managed plots in significant underreporting of employment relative to vis-à-vis female-managed plots, likely outweighs the measurement via private, self-respondent interviews, unknown risk of aggregation in many surveys, including with stronger effects for women (Kilic et al., 2020a). the LSMS-ISA. Kilic et al. (2020b) find that the common practice of proxy Due to variation in land tenure status and land use land reporting results in different findings for land assets rights, it is also important to account for seasonality in than with self-response reporting.The use of proxy report- production on plots and changes in plot management ing by the “most knowledgeable household member” when considering the unit of analysis. Depending on results in higher rates of exclusive reported and economic the agricultural season, a plot may be cultivated by a ownership of agricultural land among men, and lower rates different member of the household and use different of joint reported and economic ownership among women levels of inputs along with different cropping choices. as compared to individual, self-respondent interviews.3 Researchers have often cited asymmetries in crop type and input use, and therefore productivity and earnings, Dillon et al. (2021) also look directly at the effects of by plot manager gender. O’Sullivan et al. (2014) esti- proxy response on agricultural statistics including gender mate that, after controlling for plot size and region, dimensions of proxy response using a survey experiment productivity differences across male and female-man- with household heads, random proxies and self-reports. aged plots in Africa range from 23-66 percent. In order They find no effects of respondent type on total landhold- to appropriately account for plot-level production, ings reported for the household, but statistically significant and to enable analysis of the timing of production and effects of area cultivated by random proxy reports rela- gender asymmetries, both season and plot manager tive to self-reported land data (11 percent of the standard should be considered. The plot manager may differ deviation). Effect sizes are much larger on land reported from one season to the next, depending often on gen- der-based norms. 3 The gold standard approach of individual, self-respondent interviews for the measurement of asset ownership and control in household surveys has been recently researched and supported by the United Nations Statistical Commission through the United Nations Evidence for Gender Equality (EDGE) Initiative. The EDGE Initiative led to the publication of the United Nations Guidelines for Producing Statistics on Asset Ownership from a Gender Perspective (UNSD, 2019). 10 AGRICULTURAL SURVEY DESIGN by household heads and random proxies relative to to recall as a large expenditure would be significant to self-reports for field crops and pasture land. Household farmers. Survey designers have some discretion over the heads also over-report production of cereals, cash crops length of the recall period by choosing the timing and and crop diversity as computed by crop diversity scores, number of field visits during the agricultural year. relative to self-reports. Household heads (+ 18 percent of a standard deviation) and random proxies (- 37 percent of Gaddis et al. (2019) and Arthi et al. (2018) use methodolog- a standard deviation) also provide different biases relative ical validation studies to analyze the impact of recall on the to self-reported agricultural labor. Female proxies report measurement of agricultural labor in Ghana and Tanzania, lower levels of fertilizer for the household and higher respectively. In Tanzania, seasonal recall of agricultural labor at frequencies of crops such as legumes and vegetables that the person-plot level, a common method employed, resulted women traditionally produce in Burkina Faso. in reported labor up to four times the number of hours reported through weekly interviews (Arthi et al., 2018). SURVEY DESIGN CHOICES: Similarly, Gaddis et al. (2019) find that agricultural labor is RECALL PERIODS overestimated by approximately ten percent through the seasonal recall approach vis-à-vis weekly interviews in Ghana. Recall periods for agricultural statistics require consid- Both findings, while different in magnitude, suggest that a sea- eration of the different frequencies and seasonal timing sonal recall approach to agricultural labor measurement may of input decisions and harvest periods. The number and result in underestimated labor productivity. timing of survey visits will have immediate implications on the recall periods employed in the survey. Survey Beegle et al. (2012) estimate potential biases due to dif- visits may be employed once per year (with recall for ferences in recall periods in surveys conducted in Malawi, the full year), in two visits per agricultural season (one Kenya and Rwanda. Due to the design of these surveys at the post-planting stage to collect data on prepara- which were implemented over a 12-month period with clus- tion and planting and a second visit after the harvest, to ters randomly assigned for interview across regions, recall inquire about harvest and sales activities), or more than of agricultural input choices and harvest period outcomes two visits. The more frequent the survey visits, the less were de facto randomly assigned depending on when the recall the survey requires. We recommend the imple- household was interviewed. The authors find no evidence mentation of a two-visit approach where feasible, as this of bias in harvested quantities across the three countries, offers a balance of practical implementation feasibility examining both staple and cash crops. Malawian tobacco and limited recall bias. The reference questionnaire in farmers did over-report the cash value of their harvest, Appendix 2 assumes a two-visit approach and therefore while also under-estimating fertilizer use as recall peri- includes separate post-planting and post-harvest ques- ods increased. Households did not exhibit recall bias with tionnaires. If a two-visit approach is not feasible due to respect to fertilizer use for staple crops, except for female- budgetary or time constraints, a one-visit option may be headed households that cultivated maize. Longer recall employed. However, one-visit surveys often bear a heavy periods lowered reported fertilizer use for female-headed respondent burden as data for the full season, or multiple maize crop cultivators. Recall bias was important in hired seasons within the year, as asked about in one sitting. labor reporting, but the direction of these biases varied by country. Beegle et al. (2012) results suggest over-reporting In the context of an LSMS survey in particular, or any agri- in Kenya and under-reporting of hired labor in Malawi as culture focused survey that is also measuring household recall periods increase. Recall bias tended to differ by sub- consumption, a tradeoff in survey implementation arises group, even if overall trends were not statistically significant. when sampling designs spread survey implementation While there were no significant recall effects for Rwandan over a 12 month period to balance consumption sea- sorghum and coffee farmers, coffee farmers with small sonality. This survey implementation choice then creates landholdings did under-report hired labor as recall periods variation in the recall period for production and input increased. Using a similar identification strategy,Wollburg et information. Recall bias depends on the salience of the al. (2020) find the recall period length consistently affects event reported, its frequency and the time period over farmer reporting of plots listed, labor and other inputs, and which information must be recalled. For different types of maize production in Malawi and Tanzania, an effect that car- agricultural activities, salience can vary. For example, labor ries over to common measures of agricultural productivity. inputs from household members may be frequent during This is related to the design choice of number and timing the planting period and difficult to recall while fertilizer of field visits, as more visits and shorter roll-out periods costs incurred during the planting season may be easier reduce the length of the recall period. 2. Survey Methodology for Agricultural Data Collection 11 SURVEY DESIGN CHOICES: In even the most remote villages, a variety of information MINIMIZING ATTRITION IN based on how households and individuals are connected to LONGITUDINAL SURVEYS each other can help reduce attrition by finding respondents, but also understanding the sources of attrition. The prin- Agricultural surveys may be administered either ciple in survey design to facilitate tracking individuals and cross-sectionally or longitudinally. Longitudinal surveys, their households is to build in triangulated sources of infor- or panel surveys, offer additional benefits in terms of mation on a particular individual or household in case one monitoring change over time for the same households. source of information does not lead to success in tracking In panel surveys such as the LSMS-ISA, households (or the observation. Mobile phones are frequent in many rural other units of analysis) are tracked from survey round contexts and are useful sources of information about the to survey round. This often requires specialized fieldwork individual or household itself as well as friends who might teams that are responsible for using the information know how to contact the individual or household. collected in previous rounds and at the suspected place of residence, to locate and interview (i) households In the context of agricultural modules specifically, the that have moved from their previous residence and (ii) survey design choice depends on the researcher’s prefer- households that have “split-off” from the original survey ence for panel or repeated cross-sections of a household’s household, such as when household members leave the land portfolio. The literature on repeated cross-sections original survey households to create their own house- in estimating poverty dynamics is well-established with hold. If households are not tracked successfully, or they empirical strategies that are replicable in the context of refuse to participate in future survey rounds, this results agricultural surveys (see for example, Deaton (1985), in attrition. While panel surveys offer analytical advan- Antman and McKenzie (2007), Dang et al. (2014)). In the tages, attrition poses a risk to the representativeness of context of agricultural modules, identification of parcels the panel sample over time. and plots over time is greatly improved with the use of GPS coordinates that can help identify locations across Survey design choices can reduce attrition by facilitating surveys. However, interview costs increase considerably if tracking with high quality data, but it can also capture each parcel and plot must be visited in each survey round. information on potential reasons for household attrition While parcel or plot panels may facilitate the estimation that might be used to quantify and assess attrition bias. of dynamic agricultural modules, recall of parcels and plots The literature on attrition bias has primarily focused on across survey rounds is nontrivial. LSMS-ISA surveys have household and individual attrition. After reviewing 13 implemented parcel panels in select countries with success, national panel surveys in developing countries, Hill (2004) but have not employed plot panels primarily due to the fluid found that mobility was the greatest source of attrition nature of plot dimensions within a given parcel over time. bias. Fitzgerald et al (1998) and Witoelar (2011) provide a Approaches to address recall bias include naming parcels general review of the attrition issues in longitudinal sur- and plots, drawing maps, or taking plot characteristics from veys while Thomas et al. (2012) discuss low cost design previous survey rounds into subsequent rounds, but little protocols that limited attrition in the context of the research has been done outside of piloting to provide an Indonesia Family Life Survey. estimate of bias reduction due to different recall methods. 12 AGRICULTURAL SURVEY DESIGN 3. Production This chapter reviews agricultural survey design innova- a single crop, as multi-cropping or inter-cropping is an tions and their methodological validation studies that important land management practice to increase yield inform survey design choices related to crop production. and land quality. Third, property rights are not necessar- We focus on methodological innovations that have been ily well established in many rural contexts, and multiple validated either through survey design experiments or family members may work on the plot or cultivate the validation studies with an ‘objective’ measure. We first plot in different seasons. The land owner may be differ- discuss measurement issues related to production before ent than the person making decisions to cultivate the summarizing survey design choices and practical imple- land, as share-cropping, land leasing or land lending may mentation in the reference questionnaire provided in the mean that land owners are not making agricultural deci- appendix of this volume. sions. Holden et al. (2016) describe survey design choices in modules used to describe land tenure. Depending on LSMS surveys have primarily taken a production function the empirical applications of data collected, information design objective with complementary questions related on sources of land acquisition including inheritance and to farm revenues and costs without extensively provid- legal status, land transactions, formal and informal prop- ing a method to calculate farm profits. In early versions erty rights, land conflicts, perceptions of tenure security of the LSMS surveys and many other multi-topic house- and trust in land-related institutions may all be important hold surveys, agricultural activities were not detailed at additional questions complementary to the land roster. the plot level, as this requires a higher respondent burden This is especially true if the survey aims to monitor relative to household level recall. One innovation of Sustainable Development Goals related to land tenure, recent LSMS-ISA surveys has been to use the plot as the particularly SDG 5.a.1 and 1.4.2. Recognizing the impor- unit of analysis for field crop data collection. The advan- tance of land tenure security as it relates to investment tage of this approach is that it more accurately measures and the control of and access to other assets such as the relationship between input and outputs in agricultural credit, both of these SDG Indicators seek to measure production when the household head may not manage and monitor specific aspects of land tenure at the individ- all production. Production heterogeneity results from dif- ual level, rather than at the household level.4 ferences in crops cultivated that require different levels of inputs and management by several plot managers. The In summary, the key innovation in conducting plot-level level of detail required in subsequent modules does not production analysis is not to simply measure inputs and make this survey design choice trivial. Plot level data outputs at the plot level, but to distinguish the unit of collection requires not only measuring land area at the analysis as plot-crop-season-manager. This unit of analy- plot level, but levels of production, labor, capital, chemical sis facilitates comprehensive measurement of household inputs, and land management techniques disaggregated at production, allowing multiple analytical strategies from the plot level. seasonal, crop and gender perspectives, but also has some Choosing to measure production at the plot level requires a wider series of choices in identifying the unit 4 For more detailed guidance on data collection for monitoring SDGs of analysis which has design and empirical implications. 5.a.1 and 1.4.2, on land tenure, refer to the guidance document First, agricultural production is seasonal and may involve developed by the custodian agencies (FAO, The World Bank, and UN Habitat, 2019); available here: http://documents.worldbank.org/curated/ multiple crop cycles on a given plot that need to be en/145891539095619258/pdf/Measuring-Individuals-Rights-to-Land-An- measured. Second, plots are not always associated with Integrated-Approach-to-Data-Collection-for-SDG-Indicators-1-4-2- and-5-a-1.pdf 3. Production 13 limitations particularly in the context of a panel survey. within the survey period (such as soil characteristics, Tracking plots is difficult between survey years which is water source, plot topography) and time-varying char- one of the reasons that the LSMS-ISA surveys are con- acteristics (including labor, inputs, and capital utilization). ducted as a household panel (or household-parcel panel in We devote the next two chapters to the measurement of select countries) with repeated cross-sections of plot, pro- these time-invariant and time-variant characteristics but duction, and input information. Another difficulty is related note that allocating these inputs and labor to the plot- to plot characteristics that are relatively time-invariant crop-season-manager unit is often challenging. Table 2. Survey Decisions that Inform Production Measurement POST-PLANTING QUESTIONNAIRE Questionnaire module Key design choice Motivation/Considerations PP 1 - Parcel Roster Land area measurement Farmer self-reported land area has been found to be systematically biased so with GPS that farmer-reported land area measure is considered unreliable. Farmer- reported land area is still collected as it can be used to impute the area of parcels for which the GPS area has not been measured. Non-standard units of Reporting in standard units (square meters or hectares) may pose difficulties measurement for farmer for respondents in some cases and can worsen reporting error. Non- self-reported land area. standard units should be country-specific. In some cases, conversion factors for non-standard units may vary by region, district, province, etc. Conversion factors must be available (either through existing data or supplemental conversion factor development work) in order to convert all land areas to a common unit in the analysis phase. PP 2 - Parcel Details Individual-level ownership Asset ownership, particularly of land in agrarian settings, is a critical factor determined of individual wealth and security. Individual-level ownership of land allows for gender-based analysis and monitoring of select SDG Indicators when coupled with other questions on land. Network identification  Network identification allows respondents to identify co-ownership of parcels with individuals outside of the household. PP 3 - Plot Roster Land area measurement Farmer self-reported land area has been found to be systematically biased so with GPS that farmer-reported land area measure is considered unreliable. Farmer- reported land area is still collected as it can be used to impute the area of plots for which the GPS area has not been measured. Non-standard units of Reporting in standard units (for example, square meters or hectares) may measurement for farmer pose difficulties for respondents in some cases and can worsen reporting self-reported land area. error. Non-standard units should be country-specific. In some cases, conversion factors for non-standard units may vary by region, district, province, etc. Conversion factors must be available (either through existing data or supplemental conversion factor development work) in order to convert all land areas to a common unit in the analysis phase. Plot as primary unit of Plot-level data collection is designed to improve respondent recall and makes observation for land possible analysis of the data at the plot-level. preparation and crop planting information PP 4 - Plot Details Recording of plot Plot manager characteristics matter for plot level outcomes such as manager harvest. Identification of plot manager also allows for gender-specific productivity analysis, analysis of input allocation, gender-based crop cultivation patterns, etc. Respondent selection/ • It is recommended that the plot manager, being the most knowledgeable recording by plot person concerning the plot, respond to the questions about plot characteristics. Plot managers may differ from plot to plot which means that each plot may have a different respondent. • For data analysis, recording the ID of the respondent allows controlling for the effect that the respondent’s characteristics such as education on gender have on the data at the same level of observation of the data. 14 AGRICULTURAL SURVEY DESIGN POST-PLANTING QUESTIONNAIRE Questionnaire module Key design choice Motivation/Considerations PP 6 - Crop Roster Plot-crop as unit of  Allows for detailed accounting of each of the crops planted and enables observation for crops plot-level analysis of production. planted Permanent and The plot may contain both permanent and temporary crops so both need temporary crops in the to be reported in the same module, using the plot to anchor recall on same roster but with planting. Permanent crops, such as trees, are likely not planted in the ongoing differentiated questions. agricultural season, so year of planting instead of month of planting is recorded. Reporting of seed Reporting in standard units (e.g. kilograms) poses difficulties for respondents quantity in non-standard and can worsen reporting error. Non-standard units should be country- units specific. Reference period for The reference period aligns with the actual production cycle aiding seed acquisition is respondent recall and improving data reliability. ‘reference agricultural season’ POST-HARVEST QUESTIONNAIRE Questionnaire module Key design choice Motivation/Considerations PH 7 - Field Crop Respondent selection/ • It is recommended that the plot manager, being the most knowledgeable Production recording by plot-crop person concerning the plot, respond to the questions about harvest. Plot managers may differ from plot to plot which means that each plot may have a different respondent. • For data analysis, recording the ID of the respondent allows controlling for the effect that the respondent’s characteristics such as education on gender have on the harvest data. Reporting of harvest Reporting in standard units (e.g. kilograms) poses difficulties for respondents quantity in non-standard and can worsen reporting error. Non-standard units should be country- units specific. Reporting of harvest The weight and value of harvest may vary depending on the condition in quantity in different which it was harvested and/or weighed. conditions. Decision-maker This allows for understanding household decision-making over the concerning use of harvest production process, which is especially relevant from a gender perspective. recorded PH 8 - Field Crop Respondent selection/ Decision maker may vary from crop to crop, data considered most reliable if Disposition recording by crop each decision-maker can respond in person rather than proxy response. Reporting of harvest Reporting in standard units (e.g. kilograms) poses difficulties for respondents sales quantity in non- and can worsen reporting error. Non-standard units should be country- standard units specific. Reporting of harvest The weight and value of harvest may vary depending on the condition in sales quantity in different which it was harvested and/or weighed. conditions. Decision-maker This allows for understanding household decision-making over the concerning earnings from production process, which is especially relevant from a gender perspective. sold harvest recorded Reference period for The reference period aligns with the actual production cycle aiding seed acquisition is respondent recall and improving data reliability. ‘reference agricultural season’ 3. Production 15 LAND MEASUREMENT tation costs to individual farmer plots, and enumerator time spent measuring the plot. The use of GPS can on We briefly review the land measurement literature and average require as little as 28 percent of the time needed methodological guidance that has emerged from a recent for compass and rope (Keita and Carfagna, 2009; set of land measurement experiments, primarily focused Schøning et al., 2005). Keita and Carfagna (2009) find that in sub-Saharan African countries. Much of this litera- on small plots CR can take up to 17 times longer than ture has been summarized in Carletto et al. (2016), but GPS, though GPS measurement still requires survey team this literature continues to evolve, particularly as most relocation to the plot and survey team time to trace the empirical evidence has been focused on sub-Saharan field’s perimeter. Similarly, Carletto et al. (2017) find that African production systems. The extent of possible land CR measurement takes approximately four times as long measurement error is highly dependent on the method as GPS measurement on average, in methodological val- used to measure land size. Three common measurement idation studies in Ethiopia and Tanzania. Because of this, methods include compass and rope, plot measurement many surveys rely on a farmer’s own estimate of land size with GPS devices, and farmer self-reported land size. which avoids the time and cost of actual measurement. Remotely sensed plot measurement is also feasible However, self-reported estimates could be subject to if the GPS coordinates of the plot and its boundaries greater error given that first, many farmers in develop- are identified. A longstanding approach to collect land ing countries acquire land through informal means where size is the compass and rope (CR) method which uses record keeping (and thus information on plot dimen- poles, ropes and a compass to carry out a systematic sions) is limited and second, farmers are more likely to measure of land size (FAO 1982). With a compass, mea- give rounded and inexact size estimates. suring tape, ranging poles, a programmable calculator or other computational tool, and two to three persons, Most land measurements require the farmer to first the CR method can measure the area of a plot with sig- identify the plot correctly for the interviewer. Plot mea- nificantly more accuracy than by subjective estimates. surements are then linked to the unique plot identifier, GPS and CR measures require careful implementation but misclassification is entirely possible. Because of the by a field team, though GPS measurements may also be intensity of this field activity, monitoring systems are affected by the GPS device’s calibration error which can paramount to ensure that plots enumerated are actually be affected by environmental factors. GPS measurements visited and plot measures are properly linked to house- may be subject to more error than CR since the coordi- hold survey data, particularly if the land measurement nates measured by the GPS device are not exact.5 When teams work independently of household interview teams. carefully implemented, CR provides precise estimates, and it can therefore be considered as a benchmark Donaldson and Storeygard (2016) review the integra- against which to assess the precision of other meth- tion of satellite data for economic analysis, noting that ods. However, it should be noted that the compass and remotely sensed data were used as early as the 1930s rope method is not immune to measurement error of to study US farmland production and land use practices6. its own. In their analysis of three methodological vali- Agricultural statistics derived from remotely sensed data dation studies, Carletto et al. (2016 and 2017) illustrate have been used primarily to compare land use informa- various cases of closing error in compass and rope mea- tion and changes over time, though applications to yields surements as well as provide anecdotal evidence of the (Lobell et al., 2009) and economy-wide outcomes such challenges of this form of measurement, including signif- as food security are becoming more common (Grace et icant difficulty in the reading of compasses due to the al., 2014). Lobell et al. (2009), Grace et al. (2012, 2014), poor eyesight of many enumerators. and Husak and Grace (2016) apply remote sensing to the estimation of yields, land use, and total agricultural pro- Although the CR method is highly accurate when per- duction for a region, but are limited by their ability to link formed properly, it is time consuming and costly due to this data to specific households and other inputs in the the necessity of careful training, monitoring, transpor- agricultural production process. 5 Standard GPS devices have reported position accuracy of approximately ten meters depending on the model’s satellite calibration algorithm 6 In addition to land use information, Donaldson and Storeygard (2016) (http://www8.garmin.com/aboutGPS/). GPS receivers using WAAS note remote sensing data sources available to produce data related to (Wide Area Augmentation System) can have accuracy of three meters mineral deposits, elevation, terrain, and land cover, as well as airborne by correcting for atmospheric conditions. pollution, fish abundance and electricity use. 16 AGRICULTURAL SURVEY DESIGN While a primary advantage of remote sensing data for farmer assessment of soil in household surveys are the land measurement and land characteristics is the abil- integration of laboratory testing, which is both logistically ity to measure over large geospatial variation and easily complex and costly, the use of in-situ tools for objective construct panel measures of such variables, limitations of or pseudo-objective measures, and the integration of remotely sensed data are also described by Willis et al. survey data with available geospatial soil products. While (2015), Donaldson and Storeygard (2016), and Jain (2020). integration with publicly available geospatial soil products While remotely sensed data are not subject to respondent is likely the most cost-effective approach, recent research recall biases, satellite images are linked to GPS coordinates has found that current offerings inadequately capture vari- of farmer plots reported by farmers and images must be ation in soil properties at the local level. Berazneva et al. interpreted using either subjective evaluations or machine (2018) and Gourlay et al. (2017), for example, both suggest learning algorithms, both of which could be biased relative that the Africa Soil Information Service (AfSIS) 250m soil to actual plot size. Jain (2020) reviews errors in prepro- map (Hengl et al., 2015) fail to sufficiently capture local- cessing imagery including sensor characteristics, satellite level variation vis-à-vis ground-based objective measures. angle and atmospheric conditions.Validating classifications The future of winema measurement in household surveys of plantation, agricultural and forestry land uses, Jain (2020) likely falls at the intersection of subjective, geospatial, and find that the error structure is non-random for remotely in-situ tools. Use of a sample of ground-based soil tests sensed data with an overall accuracy of 93.9percent of land to calibrate geospatial soil products to better fit local classifications. Carletto et al. (2016) also note remotely conditions and farm management practices may result in sensed data used to measure land size may be more biased improved soil parameters at a national level, for example. for plots with larger slope gradients than relatively flat That, coupled with the growing availability and reliability of plots due to interaction with sensor angles. The data gen- soil-assessment tools, including mobile applications like the erating process to use remotely classified data requires Land Potential Knowledge System (LandPKS, landpoten- respondent input, measurement from remote sensors, and tial.org) and handheld spectrometer technology suggests the processing of this information, linking data sources for improved soil data in household surveys is soon to be real- interpretation and analysis. Remote sensing is not without ized, though additional methodological research is needed. potential implementation biases. Remotely sensed images must be relevant to the validation period’s season and year. Methodological studies on land area Imagery must be of high resolution to facilitate respon- measurement bias dent identification of plots. Respondents must be capable of identifying their plot on an image which may be com- The spread of GPS measured land size has led to sev- plicated by changing plot boundaries or ambiguous plot eral studies that examine the differences between the boundaries. For these reasons, despite the advantages of GPS and both CR and self-reported measures including remotely sensed data, empirical applications using remotely the potential impact of measurement error resulting sensed data should also be careful of potential biases. from inaccurate land size measurements. In two examina- tions of GPS versus CR measures in sub-Saharan Africa, Land quality, in addition to quantity, is essential for produc- Schøning et al. (2005) and Keita and Carfagna (2009) find tivity measurement yet challenging to measure accurately that GPS based measurements are generally lower than in a household survey context. Household survey instru- CR measures though neither study provides a technical ments, including the one included in Appendix 2, typically explanation for this difference. Both studies find that the collect information on soil quality and health through a GPS-CR difference is statistically significant for smaller series of subjective farmer assessments. Findings from plots but not for larger plots7. It appears that for smaller methodological research, such as that by Carletto et al. plots, the fixed margin of error associated with deter- (2017a) and Gourlay et al. (2017), illustrate the limitations mining GPS coordinates is large relative to the size of of such subjective soil data, highlighting the limited vari- the plot and thus results in a greater measurement error. ability observed in such data and the weak relationship of certain subjective questions, particularly overall soil quality, with laboratory-based soil quality indicators. This finding 7 Schøning et al. (2005) find a statistically significant difference for plots is echoed by Berazneva et al. (2018), who find in Kenya under 0.5 hectares but not for plots equal to or greater than 0.5 hectares. that farmer-reported soil type is reasonably correlated Keita and Carfagna (2009) separate their sample into five clusters based on plot size. They find a statistically significant difference for the clusters with objective measures of soil fertility while farmer-re- with the lowest land size but not for the clusters with larger plots. ported soil quality is not. The alternatives to subjective However, they do not specify the plot size range for each cluster. 3. Production 17 Since most plots in developing countries are relatively mate for medium size plots. In a similar analysis, Carletto, small, the potential for land mismeasurement using GPS Gourlay, and Winters (2015) find inconsistencies between could be significant. However, rounding of very small plot GPS and self-reported measurements in Malawi, Uganda, areas in Schoning et al. (all plots less than 0.01) clouds the Tanzania, and Niger. They find that in all four countries reliability of their findings for smaller plots. self-reported plot sizes were higher than GPS for smaller plots but lower for larger plots, suggesting that farmers A cross-country study by Carletto et al. (2015) showed over-report the area of small plots and under-estimate the that the error in farmer-estimated area is both signifi- area of large plots. Carletto, Savastano, and Zezza (2013) cant in magnitude and systematic in nature. Data pooled find that in addition to plot size, the rounding of self-re- from nationally-representative surveys in Niger, Malawi, ported measurements, the age of the household head, and Tanzania, and Uganda suggests that self-reported areas whether the plot was in a dispute with relatives were all are underestimated by 2.5 percent relative to GPS mea- positively associated with a greater GPS-self-report differ- sures on average, but that respondent estimation on the ence in land size in Uganda. Carletto, Gourlay, and Winters smallest plots is overestimated by more than 100 per- (2015) similarly find that rounding of farmer estimates are cent on average while respondents underestimate the consistently associated with a difference between GPS area of larger plots (Carletto et al., 2015). Findings from and self-reported plot size in all four countries. Dillon et methodological studies in Ethiopia, Tanzania, and Nigeria al. (2019) also find that GPS measurements of land area illustrate a similar trend, with self-respondent area esti- are similar to compass-and-rope (CR) estimates and more mation significantly overestimated on the smallest plots, reliable than farmer estimates, where self-reported mea- with overestimation decreasing with plot size (Carletto surement bias leads to overreporting land sizes of small et al., 2017). Findings from pooled data from methodolog- plots by 83 percent and underreporting of large plots by ical studies in Ethiopia, Tanzania, and Nigeria, however, 21 percent of the compass and rope estimate. Their study suggest that on average GPS plot area measurements in Nigeria finds that the error observed across land mea- are only 1 percent different that the CR measurements surement methods is nonlinear, is not resolved by trimming (Carletto et al., 2017). On the smallest plots, those under outliers, and results in biased estimates of the inverse land 0.05 acres, the mean GPS and CR measures are not sta- size–productivity relationship. A key econometric advance tistically different (Carletto et al., 2017). Similarly, Desiere in this paper is the ability to control for plot fixed effects (2015), in an analysis of more than 50,000 agricultural which may bias parameter estimates. They also investigate plots in Burundi, finds support for implementation of input demand functions that rely on self-reported land GPS measurement on small plots.The study finds that, for measures and find that these measures significantly under- plots smaller than 500m2, bivariate regressions of GPS on estimate the effect of land on input utilization including CR explained approximately 97 percent of the variance, fertilizer and household labor. though they do cite increased precision of GPS measure- ments as plot size increases (Desiere, 2015). Dillon et al. (2021) estimate measurement differences between self-reported, GPS and remotely sensed land Several studies have found inconsistencies between GPS size measures. In their comparison of results from Lao and self-reported land sizes (Goldstein and Udry 1999; People’s Democratic Republic, Philippines, Thailand and Carletto, Savastano and Zezza, 2013; Carletto, Gourlay Viet Nam, they find significant differences between GPS and Winters, 2015; Dillon et al., 2019; Dillon et al., 2021). and remotely sensed data only in Viet Nam, where plot Although ancillary to their main analysis, Goldstein and sizes are small relative to the other countries. The mag- Udry (1999) found that the correlation between GPS and nitude of farmers’ self-reporting bias relative to GPS self-reported land size was only 0.15 in their dataset from measures is nonlinear and varies across countries, with Ghana. The authors do point out that, historically, field the largest magnitude of self-reporting bias of 130 per- measurements in the region were based on length and not cent of a standard deviation (2.2-hectare bias) in the Lao area and that this could partially explain the lack of a strong People’s Democratic Republic relative to Viet Nam, which correspondence between farmer and GPS estimates. has 13.3 percent of a standard deviation (0.008-hect- However, Carletto, Savastano, and Zezza (2013) also find a are bias). While the land measurement literature has sizeable difference in Uganda though one that varies with expanded largely due to LSMS-ISA collaborations in plot size.The authors find that the bias is wider for smaller Africa, variation across cropping systems in other con- (less than 1.45 acres) and larger (greater than 3.58 acres) tinents can yield different types of land measurement plots while self-reported measures are a reasonable esti- biases in self-reported land data. 18 AGRICULTURAL SURVEY DESIGN Table 3. Research on Land Area Measurement PAPER LAND COUNTRY/DATASET KEY INSIGHTS MEASUREMENT METHODS Goldstein and Udry Farmer self-report; GPS Eastern Ghana / Agricultural • Correlation between self-reported and GPS (1999) Innovation and Resource measured plot size is only 0.15 Management in Ghanaian • Proposed interpretation: agricultural Households survey history in Eastern Ghana in which local field measurements were often based on length rather than area, making it hard for respondents to report in square meters De Groote and Farmer self-report; Southern Mali / Field experiment • Relative to compass and rope, farmer self- Traoré (2005) Compass and Rope data collected by Farming Systems reported area (supported by enumerators) Research team in Mali over-estimated on small plots, under- estimated on large plots, but under- reported by 11 percent on average. • Measurement error smaller for cotton fields than for cereals Schøning (2005) Farmer self-report; GPS; Uganda / 2003 Uganda Pilot • Difference between GPS and Compass and Compass and Rope Census of Agriculture Rope is small and significant on small plots (<0.5ha) but not on larger plots. • Farmer self-report unreliable. • Compass and Rope takes 3.5 times as long as GPS. Keita and Carfagna GPS; Compass and Rope Cameroon, Madagascar, Niger, • GPS tends to underestimate plot size (2009) Senegal / sample of 207 relative to Compass and Rope, but the purposively selected plots from error is small. Cameroon, Madagascar, Niger, • Accuracy of GPS appears to be higher Senegal without cloud cover. • No correlation between plot size and difference between GPS and Compass and Rope measurements • Compass and rope takes on average takes 3.8 times longer than GPS, and this is worse for medium and small plots. Carletto et al. Farmer self-report; GPS Uganda / Uganda National • Relative to GPS, land area over-estimated (2013) Household Survey 2005/06 on small plots, under-estimated on large plots. • Other determinants of measurement error: rounding, boundary delineation of plot, tenure status of plot Carletto et al. Farmer self-report; GPS Malawi, Niger, Tanzania, Uganda • Relative to GPS, self-reported area of small (2015) / LSMS-ISA Malawi Integrated plots (less than 0.5 acres) systematically Household Survey 2010/11, over-estimated. Uganda National Panel Survey • Degree of over-estimation varies, but in all 2009/10, Tanzania National Panel countries the mean self-reported area is Survey 2010/11, Niger ECVM/A overestimated by at least 90 percent of the 2011 mean GPS area of plots on small plots. • Area of large plots under-estimated Desiere (2015) GPS; Compass and Rope Burundi / Enquete nationale • GPS slightly over-estimates area relative to agricole du Burundi 2011/12 Compass and Rope area but the under- estimation is negligible • GPS is reliable even on smaller plots (less than 0.5 hectares), with reliability increasing rapidly for plots larger than 0.1 hectares. 3. Production 19 PAPER LAND COUNTRY/DATASET KEY INSIGHTS MEASUREMENT METHODS Carletto et al. Famer self-report; GPS; Ethiopia, Nigeria, Tanzania • Error in self-reported area depends on plot (2016 and 2017) Compass and Rope (Zanzibar) / Ethiopia Land and area, with small plots over-estimated, large Soil Experimental Research plots under-estimated. 2013, Nigeria Area Measurement • In comparison to Compass and Rope, GPS Validation Study, Measuring is generally very reliable. Cassava Productivity study Tanzania (Zanzibar) Kilic et al. (2017) Famer self-report; GPS Uganda, Tanzania / Uganda • Not all plots are measured with GPS National Panel Survey 2009/10, because some are too far away or Tanzania National Panel Survey otherwise inaccessible, which may lead to 2010/11 bias. • Farmer self-reports are predictive of GPS measured area so that Multiple Imputation techniques can be used to impute area of plots where GPS is missing. Dillon and Rao Farmer self-report; GPS; Lao PDR, Philippines, Thailand, • Remotely sensed plot area differs from GPS (2021) satellite data Vietnam / land measurement area estimate only in Viet Nam, where plot experiment among rice producers sizes are small. • Farmer self-reported plot area estimates are most overstated for small plots. Dillon et al (2019) Famer self-report; GPS; Nigeria / Nigeria General • GPS more reliable than self-reporting, Compass and Rope Household Survey 2012/13 relative to Compass and Rope • Farmers underestimate land area on large plots and overestimate it on small plots. • GPS error is relatively small relative to Compass and Rope. FIELD CROP PRODUCTION AND data. Conversions require not only question response YIELD MEASUREMENT: SELF-REPORTED, category flexibility to allow respondents to report quan- REMOTE-SENSING, AND CROP-CUT tities in non-standard units, but inter-linked data collected PRODUCTION MEASURES through market surveys to measure systematically con- versions of non-standard units by crop or product into Accurate measures of production likely rely on reducing standardized units (kilograms or liters). Alternative the cognitive burden of reporting for the respondent. approaches to collecting non-standard unit conversions As applied to the measurement of production, plot- formally in the market survey is to design a module to level recall by plot managers may actually be cognitively ask respondents directly for conversions, but a respon- less burdensome than asking a respondent, potentially dent’s estimate of the conversion factor may not be made the head of household, to aggregate across all house- based on objective standards, creating bias in measures hold plots to report production. Plot level production across respondents that is directly correlated with their reporting also allows women’s production to be more conversion estimation skill. A significant benefit of CAPI accurately measured if intrahousehold gender asymme- implemented interviews is the possibility to document tries exist. In facilitating respondents’ accurate reporting, in advance non-standard measures which provide easier a principle of agricultural survey design has been to let reference for respondents, as well as a pre-loaded con- respondents report in the units in which they have mea- version factor. A classic example is bunches of bananas sured their own production. This principle opens the for which standardizing even this non-standard measure possibility that respondents may not use standard units is particularly challenging. of measures, such as kilograms to report production. Alternatives to self-reported production include more Oseni et al. (2017) present a comprehensive guide for intensive field approaches such as crop-cutting or inte- measuring conversions from non-standard units to stan- grated survey techniques such as first geo-referencing dard units for consumption and agricultural production plots and then estimating production using remotely 20 AGRICULTURAL SURVEY DESIGN sensed data, or remotely sensed data in combination evidence from the land measurement literature that find with ground-based data. It is important to underscore nonlinear plot reporting effects, particularly for smaller that using either crop-cutting or remotely sensed data plots. Gourlay et al. (2019) find similar results in the to measure production involves estimation of production analysis of maize in Uganda, with self-reported yields quantities. While these approaches may be more reli- being significantly overestimated on the smallest plots able than self-reported production by farmers, they are (by more than 3,500kg/ha on average in the smallest not without bias themselves. Researchers should avoid plot area quintile). A divergence between self-reported labeling these measures as objective or direct measures and crop-cut-based yields is also observed by Lobell et unless the full plot is crop-cut because plot samples are al. (2020) who find that, although the means are sim- used to then extrapolate total plot production. ilar, self-reported estimates of sorghum yields in Mali are only weakly correlated with yields measured via Crop-cutting methods of plot production estimation and crop-cutting (correlation coefficient of 0.33). Bevis and yield calculation can be undertaken using several different Barrett (2020) argue that agricultural labor intensity measurement methods. Plots are defined by field teams increases at a plot’s edge and thus drives erroneous and respondents to delineate the plot’s size. Plot size is production self-reports and the inverse land size rela- measured often using either compass and rope or GPS tionship in their paper which also uses crop-cuts to techniques. The plot area is demarcated, and a crop-cut- measure production. Abay et al. (2019) estimate the ting rule is implemented where the field area is sampled, magnitude of bias on the inverse land size relationship the sampled plot areas are clear cut, production is dried, describing potential econometric implications of cor- weighed, and then based on the sampling fraction of the recting both land size and production measures. An plot samples, total production is estimated. Crop-cuts are important insight from econometric applications that estimates of crop production, potentially more reliable they apply to agricultural data is that correcting one than self-reports, but subject to biases from the measure- measurement error in either land or production esti- ment process. First, budgets will determine the sampling mates may increase bias when measurement errors fraction of plot subsamples taken from the plot from are correlated. They find that reducing bias by using which the plot production estimate will be calculated. crop-cutting production measures with compass and The higher the sampling fraction, the more reliable the rope land measures causes the inverse land size rela- plot estimate. Second, crops must then be dried, threshed tionship to disappear in their Ugandan sample. and weighed which often takes additional days of field- work, sample tracking protocols, and standardization of The evidence on remote sensing to measure production drying and weighing techniques. Finally, the plot sub-sam- is an evolving scientific field. Carfagna and Gallego (2005) ple production weights are then used to extrapolate the argue that remotely sensed data are particularly useful total plot production measure. Crop production may vary for area frame measurement to define sampling units but within plot due to planting techniques, plot slope that hesitate to recommend remotely sensed images to esti- may pool water differentially within the plot, uniformity mate crop production because the classification of pixels of plot fertilization and weeding, or soil quality. Animal assigned to specific crops is often strongly biased. This is infestation or weather damage may also not be uniform one of the reasons remote sensing-based estimation of within the plot. Fermont and Benson (2011) review the yields in intercropped areas is particularly challenging, as sources of bias and history of crop cutting in Ugandan discussed in Lobell et al. (2019). This field continues to agricultural surveys, noting the widespread variation in evolve and identification algorithms linked to biological crop cutting measurement methods and yield estimates science are improving. It is important to remember that using various crop-cutting measures. measurement of plot production is based on models of interpreting visual images that correspond to biological A large body of literature has focused on comparisons plant growth. Brown and Pervez (2014) provide evidence of crop-cut measures relative to self-reported mea- in the US context that some models of production pre- sures, particularly focused on the empirical relationship diction are increasingly accurate to measure land use between yield and land size. Desiere and Joliffe (2018) and production, calibrating and validating their models find that self-reports over-estimate yield on small plots with USDA data. In Uganda, Lobell et al. (2019) estimate and underestimate yield on large plots relative to the maize plot yields using (i) full plot crop cuts, (ii) partial, crop cut measures which they use as their validation sub-plot crop-cutting, and (iii) remotely sensed images of measure. Their production results largely corroborate plot yields. To estimate yields via remote sensing imagery, 3. Production 21 Lobell et al. (2019) estimate two models, one calibrated increased annual production reports by 28 percent com- on ground-based, crop-cutting data and one without pared to the in-person supervised diary visits. calibration of ground data. They find that calibrated remotely sensed yield estimates captured half the vari- We have also sidestepped an important complication in ability of yield estimates in comparison to full-plot crop crop production measurement, specifically the case when cut production measures for plot sizes above 0.1 hectare. production systems use mixed or inter-cropping planting Uncalibrated remotely sensed yield estimates were one techniques. LSMS-ISA data from Tanzania, for example, ton per hectare higher than crop-cut measures. These indicates that approximately 64 percent of cultivated results are both incredibly encouraging and require a plots are cultivated with more than one crop (Wineman further research agenda to better understand best prac- et al., 2019). Estimating both land area apportioned to tices for remotely sensed yield estimates that extend to a a particular crop and its production is particularly chal- wider set of crops. lenging. Measuring yields on these intercropped plots can take many forms, through varying methods of esti- Challenges to production measurement mating the denominator, including: (i) using the full plot in intercropped systems area for each crop; (ii) using the share of plot area under a given crop, identified by seeding rate, plant density, Much of the literature on production measurement has or area estimation, such that the total cultivated area considered the case of a single field crop, but farming sums to the total plot area; (iii) allocating total plot area systems are often more complex and not all agricultural equally across crops cultivated on the plot such that the production is based on field crops. We review measure- total cultivated area sums to the total plot area; or (iv), ment approaches when planting techniques on a plot use imputing the area which a crop would occupy if it were mixed, multiple crops or inter-cropping techniques as well mono-cropped (where the total cultivated area summed as cases of roots, tubers and tree crops whose harvest across crops may exceed the total plot area). Most period may be multiple rather than at a singular period household surveys acknowledge the complications of of time in the production process. The LSMS-Integrated production and input estimates on inter-cropped plots by Surveys on Agriculture incorporate seasonal planting and identifying these plots and apportioning the area planted harvest for annual crops in its biannual interview struc- to divide plot-level input reports to production reported ture, but concern over recall bias for root, tuber and by crop. Proportional input attribution implies crop input tree crops that do not overlap with the main agricultural demands including fertilizer, weeding, and harvest time season varies across agro-ecological zones and across are similar by crop. national surveys. Production surveys focused on root, tuber and tree crops are often timed to the main har- The Global Strategy to Improve Agricultural and Rural vest periods for these crops, but supplement production Statistics (2018) provides methodological guidance on reports with higher frequency surveys. Kilic et al. (2018) implementing the above methods to measuring the area conduct a survey experiment in Malawi over a 12-month under a given crop in intercropped systems. However, period where variations of diary and farmer recall are a best practice recommendation supported by a meth- used to produce cassava production estimates. Their pro- odological survey experiment is not currently available. duction reporting treatments vary the recall period (daily Remote sensing or crop-cut production estimates are diary recalls (supervised in person or via telephone), possible alternatives, but these measures are also chal- six-month and annual recalls).The survey experiment indi- lenging to implement. Lobell et al. (2019) report lower cates that annual recall underestimates annual production accuracy of remotely sensed production estimates by 21 percent relative to the in-person supervised diary compared to crop-cut production estimates for maize visit. Daily diary reporting supervised by a telephone visit intercropped plots in Uganda. 22 AGRICULTURAL SURVEY DESIGN Table 4. Research on Production Measurement PAPER PRODUCTION CROP/COUNTRY/DATASET KEY INSIGHTS MEASUREMENT METHODS Fermont and Various, especially Various/Various, focus on • Crop cutting associated with over-estimation of Benson (2011) farmer self-report vs Uganda / Various production production in many contexts crop cutting datasets, Agricultural Censuses • Farmer estimation of production also exhibits from 1965, 1990-91, 2008-09; significant biases; appear to be closer to objective Annual Survey from 1967-68, measures in several cases. Uganda National Household • Lack of systematic evidence on sources and effects of Survey 1999/2000 and 2005/06 biases make recommendation impossible Deininger et al. Farmer self-report; Various/Uganda / Uganda • End-of-season recall diverges significantly from (2012) harvest diaries National Household Survey harvest diaries, which are deemed more reliable. 2005-06 • Findings vary by crop type: for most crops, end- of-season recall associated with under-reporting, especially for extended-harvest crops such as cassava or banana; cash crop production recorded as significantly higher in end-of-season recall module. Kilic et al. Farmer self-report; Cassava/Malawi / • Comparison of weekly harvest diaries to two (2018) harvest diaries; crop Methodological Experiment on different recall methods for cassava production in cutting Measuring Cassava Production, Malawi: a single visit with a 12-month reference Productivity, and Variety period and two visits with 6-month reference Identification 2015-16 periods. • Relative to crop cutting, the recall methods lead to under-reporting, especially with 12-month reference period. • Crop cutting is an upper bound for extended harvest crop cassava. • Harvest diaries, especially with the support of mobile phones, perform well to capture cassava production. Desiere and Farmer self-report; Maize/Ethiopia / Ethiopia • Relative to the crop cut benchmark, respondents Jolliffe (2018) crop cutting Socioeconomic Survey 2011-12 over-report maize production on small plots and and 2013-14 under-report it on large plots. Gourlay et al. Farmer self-report; Maize/Uganda / • Relative to the crop cut benchmark, respondents (2019) partial and full plot Methodological Experiment on significantly over-report maize production on smaller crop cutting Measuring Maize Productivity, plots. Soil Fertility and Variety 2015 • Over-reporting is related to rounding/heaping and 2016 Abay et al. Farmer self-report; Wheat/Ethiopia / Randomized • Relative to crop cutting, self-reporting overestimates (2019) crop cutting controlled trial of 482 wheat production, with small plots overestimating more farmers in Ethiopia 2013-14 than large plots. Bevis and Farmer self-report Various/Uganda /IFPRI, • The edges of plots are more productive than the Barrett (2020) UBOS, and National Science inside, which may explain why farmers overstate yield Foundation panel survey of 972 on small plots and understate yield on large plots and households, 2002-3 and 2013. may also affect reliability of crop cuts. Lobell et al. Remote sensing; Maize/Uganda/ Methodological • Self-reported performed very poorly relative to full- (2019) farmer self-report; Experiment on Measuring plot crop cut. crop cutting; full plot Maize Productivity, Soil Fertility • On average, difference between partial crop cut and crop cutting and Variety 2015 and 2016 full-plot crop cut not significant. • Crop cut yield estimates captured one-quarter of full crop cut yield variability. • Both calibrated and uncalibrated satellite yield estimates captured half of full crop cut yield variability on pure stand plots above 0.10 hectare. • Uncalibrated yield estimates were consistently one ton per hectare higher than full-plot and partial crop cut. 3. Production 23 AGRICULTURAL PRODUCTION DATA: of the crop was harvested, and self-reported production DESIGN FEATURES OF THE information. If the harvest has not been completed, infor- REFERENCE QUESTIONNAIRE mation is collected on when the harvest is expected to be completed and how much of the harvest had been com- We explain the questionnaire structure in the post-plant- pleted to date. Finally, information on who controls the ing and post-harvest reference questionnaires that relate disposition of the crop harvest is recorded by respondent to measuring production that are reviewed in this chap- ID. The disposition of field crops is focused on capturing ter and earlier summarized in Table 2 “Survey decisions unprocessed crops sales including the quantity and total that Inform Production Measurement.” To measure pro- value of such sales. The volume of transactions related duction and farm productivity using yields, the reference to the total sold and the transportation costs associated questionnaire first enumerates the household’s par- with the crop sales are also recorded. Other dispositions cels in the post-planting questionnaire using the parcel of field crops include auto-consumption, gifting and in-kind roster and parcel details which includes the parcel’s size payments. Lastly, the storage of crops is recorded includ- and tenure status. A parcel is defined as a piece of land ing the quantity and method of storing. The intended use exploited by one or more persons as a single farming of the storage for later sale, future consumption, use as unit. A parcel may be bounded by natural boundaries seed or animal feed are asked in this module. and may comprise one or more plots. After parcels are listed, plots within parcels and their characteristics are We make special note in this section of the questionnaire recorded in the plot roster and plot details. A plot is on the calculation of output prices from the sales data defined as a continuous piece of land on which a unique recorded in these modules. Revenues recorded provide the crop or a mixture of crops is grown, under a uniform, basis for the calculation of output prices, but special note consistent crop management system. All outputs and of the location of sale is important analytically. In theory, inputs are recorded at the plot level. farmgate prices should reflect the output price from sale on the farm or local market, yet crop sales often occur in The primary differences in the parcel and plot details mod- markets outside of the farmer’s village. Revenue from sales ules of the questionnaire relate to the acquisition of the in larger markets outside the farmer’s village may include parcel and its tenure status in the parcel details module, transportation cost premiums above the actual farmgate and the management and soil characteristics of the plot price. The timing of sale may also affect imputed output including water sources and soil type in the plot details prices from revenue data. A large literature explores the module. The crop roster in module 6 documents at the seasonality of crop prices which may reflect risk or stor- parcel-plot-crop level the list of crops cultivated on the plot age premiums in addition to the farmgate price. From this during the reference agricultural season and seed varieties. discussion, it is also clear that not all revenue sales repre- The module also includes a question about productivity sent the ‘market’ price as a market survey might collect. expectations which can be used to compare with actual In the literature on food prices imputed from expendi- productivity recorded in the post-harvest questionnaire. ture surveys, Deaton and Zaidi (2002) recommend taking the village or enumeration area median price to impute In the post-harvest questionnaire, the parcel and plot consumption per capita to account for variations in food roster can be updated in modules 1 and 2. The plot- prices. In a similar sense, estimating output prices from crop roster in module 6 validates the crops that were agricultural modules ideally would account for sale loca- reported to have been planted from the post-planting tion as well as the timing of sale. questionnaire and records information on why crops may not have been harvested. The primary sections of the The tree and permanent crop production module is orga- questionnaire related to production are found in modules nized in a different manner than the field crop module. seven to ten. Modules seven and eight record information The primary motivation for this difference is that the on production and the disposition of field crops, while number of trees per plot is necessary to estimate yield modules nine and ten capture production and disposition for tree crops.The last harvest date is recorded as well as of permanent tree crops. the quantity of production and any losses on the parcel- plot-tree/permanent crop unit of analysis. The tree and The field crop production module includes information permanent crop disposition module is similar in structure on field crop harvest including when the harvest of a to the field crop disposition module, but uses a 12-month particular parcel-plot-crop unit began, what percentage recall period. 24 AGRICULTURAL SURVEY DESIGN 4. Agricultural Inputs The production function approach to questionnaire or other inputs. Researchers and survey practitioners design requires mapping inputs to the outputs produced interested in measuring participation rates in work and at plot level. This chapter reviews the measurement employment are encouraged to consult the 19th Inter- of chemical and organic inputs, agricultural labor and national Conference of Labour Statisticians (ICLS) agricultural capital allocated to the plot including farm standards and the pilot studies conducted by the Interna- implements and machinery. The primary measurement tional Labour Organization related to the measurement challenge in designing an agricultural questionnaire to of those concepts (Benes and Walsh, 2018a, 2018b). The capture seasonal plot level production is precisely the focus in this section is exclusively on agricultural labor attribution of inputs to the plot. Agricultural labor is conducted on the household farm, including that by drawn from the household’s members, exchange labor, or household members and hired and exchange laborers.8 hired labor from the community. In the case of exchange and hired labor, recall at the plot level may be facilitated Sagesaka et al. (2019) summarizes much of the current by supervision requirements, the timing of the labor literature on measuring work in agricultural household demand during the production process (planting, harvest, surveys. Dixon-Mueller and Anker (1988) highlight how etc.) and the relative infrequency of these types of labor the classification of workers based on their main activ- on plots. Household agricultural labor recall may suffer ity leads to underestimating the number of economically from attribution of household labor across multiple plots active women, primarily because women contribute and over relatively longer recall periods. These measure- significantly to both household and agricultural work. ment challenges are closely related to general challenges A significant number of survey experiments have also in time use modules. Fertilizers, herbicides and pesticides addressed survey design issues in agricultural labor are normally purchased in bulk and distributed across measurement. A key set of survey design choices in agri- the household or farmer’s plots. Measuring machinery cultural labor measurement have been how to frame or farm implement use at the plot-level is related to the questions regarding farm work and whether labor infor- attribution challenge of non-labor inputs. Machinery and mation needs to be self-reported or permissive of proxy farm implements are capital goods that are used over response. Bardasi et al. (2011) estimate biases in labor multiple agricultural seasons. Quantifying how much they force participation, hours worked and income by gender are used on a particular plot and the associated cost and sector of employment due to questionnaire design depends on the type of machinery and agricultural pro- such as screening questions and proxy response. In their duction process. survey experiment from Tanzania, they find no difference in female labor statistics due to proxy response, but AGRICULTURAL LABOR lower male employment rates due to underreporting agricultural activity by proxy respondents. In Malawi, Kilic A large body of literature is devoted to agricultural labor et al. (2020) find that proxy reporting increases underre- measurement to understand national employment, labor porting of employment when recall periods increase and productivity and worker earnings. The primary moti- when women are the subject of proxy reporting. Arthi vation for agricultural labor measurement is to better et al. (2018) estimate the effects of recall period, either understand smallholder agricultural productivity and its determinants. One of the most important of these 8 For guidance on the collection of data for labor more broadly, see the determinants being is labor for which poor households LSMS “Guidebook on Labor: Work and Employment in Multi-Topic have relatively larger endowments compared to capital Household Surveys” (forthcoming). 4. Agricultural Inputs 25 weekly agricultural labor reporting or end of the season Dillon et al. (2012) extend the analysis from Bardasi et al. reporting. End of season recall increases by four times the (2011) in the context of child labor statistics and survey hours reported by individuals at the plot level relative to design, finding that in rural Tanzania proxy reporting those reporting weekly. They note aggregation to house- from adults did not reduce child labor reporting rela- hold level reporting causes the reported hour differences tive to children’s own self-reports. Rather, classification between the weekly and end of season recall periods to using less precise questions regarding work activities that disappear. Arthi et al. (2018) interpret these findings as should be classified as labor force participation reduced evidence that recall bias is driven by mental burdens of by 16 percentage points girl and boy reported participa- reporting such disaggregated time use patterns as well tion in agriculture in comparison to survey designs that as memory. Given that these biases may differ depend- included screening questions listing specific work activ- ing on the level of aggregation at which they are used, ities. The Tanzanian results are somewhat surprising in the Tanzanian experiment suggest that agricultural labor that the presumption of the international community is productivity would be understated which are similar to often that parents will want to avoid reporting child labor. findings by Gaddis et al. (2019) in Ghana. However, in rural communities such as those in Tanzania, child work in agriculture is not uncommon and social Given the difficulties of measuring labor productivity, stigma is low as children’s work is viewed as human capi- due to attributing output to a particular worker and the tal building. Careful training of interviewers to be mindful unit of time in which they produced such output, Akogun of cultural context may help improve reporting on (pre- et al. (2020) measure the physical activity of sugarcane sumed) sensitive topics. cutters which is a direct measure of effort in their piece rate wage setting. They find a high correlation between SEED, FERTILIZER, PESTICIDES administrative data on output per worker recorded by AND HERBICIDES the firm and the worker’s physical activity, as well as large changes in the intensity of such activity in response Measuring seed acquisition in input modules requires a to malaria testing and treatment. Integrating physical questionnaire design that differs from recording other activity measures into national surveys may be possible non-labor agricultural inputs. This is primarily due to using a subsample to calibrate biases in reported time the common informality of seed exchange among rural use as well as predict effort-based measures of agricul- producers, but even in more formal seed markets, sig- tural labor productivity. nificant differences in seed brand names and seed traits lead to potential misclassification. For some empirical An often overlooked agricultural labor measurement applications, the researcher may want to link seed brand issue is the classification of children’s time in agriculture. names to consumer choices. In many seed applications, The ILO’s International Conference of Labour Statis- the researcher may be interested in the complementary ticians in 2008 focused on child labor measurement nature of seed choice and other input choices, or the methodology and statistics. ILO (2008 and 2017) sum- return to specific seed traits such as drought tolerance marizes many of the measurement issues related to the or germination speed, in which case, the seed trait is the classification of children’s time in agriculture as work, child relevant seed characteristic. These classification issues, labor, or hazardous work. Definitions of child labor and identifying both brand and traits of a specific input, are hazardous work depend on the country legal and agricul- not exclusive to seeds. In the case of fertilizer, pesticides tural production context, despite efforts to standardize and herbicides, brand names are not directly aligned to official reporting of child labor statistics. Guarcello et al. chemical compositions of the input. Fertilizer, pesticide (2008) review 87 datasets from 35 countries that col- and herbicide packaging lists the chemical composition lect information on child labor. They conclude variation of the product which can be used to identify product in child labor statistics varies more broadly within and traits along with the brand name. Non-standard unit across surveys relative to other children’s activities such reporting and planting technique or application method as schooling. While variation such as questionnaire design are also important non-labor agricultural input mea- and sampling do explain some of this variation, Guarcello surement issues. We have discussed non-standard units et al. (2008) note the importance of interviewer training of reporting in the previous chapter. We note Oseni et and social stigma as potentially influential factors that may al. (2017) provides extensive discussion of measurement also create variation in reported child labor information. techniques for non-standard units. We also highlight that 26 AGRICULTURAL SURVEY DESIGN planting techniques and application methods are criti- their sample. Michaelson et al. (2018) report on farmer cal to understand as complementary to the traits of the perceptions of input quality, comparing farmer per- non-agricultural inputs that agronomically will affect the ceptions with nutrient quality from laboratory-based returns to these inputs. Planting and land-management measures. They find that in their sample from Tanzania choices such as micro-dosing, distance between crops, fertilizers purchased in the market were not adulter- mounding or bunding, or inter-cropping will change the ated, yet farmers believed that they had been based on potential yield of seeds. Mixing techniques for herbi- observable characteristics of the fertilizer sample which cides and pesticides, as well as application methods affect are not necessarily correlated with nutrient quality. chemical absorption. Similarly, fertilizer application meth- ods such as broadcasting and micro-dosing will affect the Machinery and Farm Implements return to these inputs. Agricultural capital in the form of machinery and farm An objective of measuring input use is to understand implements can increase the productivity of smallholder the returns to different agricultural investments, hence farmers. Understanding how farm size and profitability measuring the often-unobservable quality of inputs is change over time are linked to the mechanization of an important characteristic of input investments. With agriculture. While it is generally regarded as easy for many products, quality is not completely observable and farmers to recall agricultural capital within the house- quality perceptions matter. This holds true for seeds hold, the plot-level attribution and control of such and other non-labor inputs. Improved seed varieties capital are measurement challenges. Plot level attribution have been shown to produce greater, more resistance of machinery use is often avoided as it may be assumed yields. However, due to several factors including seed by the survey designer that agricultural capital is shared recycling practices, informal or unregulated seed mar- equally in the household. kets, counterfeit seed, and asymmetric information, among others, farmers are often uninformed or misin- A large literature on women’s empowerment in agriculture formed of the true variety they are cultivating. Given has focused on the correct measurement between women the numerous stages of the seed supply chain in which and men’s ownership of assets versus their use-rights. variety can be masked or modified, it is not surprising Alkire et al. (2013) and Doss and Kieran (2014) provide that farmer knowledge of seed variety is generally lim- a comprehensive review and guidelines for collecting ited. Farmer identification of simply the type of seed, sex-disaggregated asset data which apply generally to agri- whether improved or local seed rather than the spe- cultural capital modules. They emphasize the importance cific variety, is still challenging. Wineman et al. (2020) of respondent and the method of collecting ownership and find, from a methodological study comparing farmer use-rights. Data from the machinery and farm implements seed type identification against DNA fingerprinting, modules can be linked to plot disaggregated production that farmers in Tanzania correctly identified maize seed and other inputs modules to assess differences in intra- type (improved or local) 70 percent of the time. Sim- household allocation of inputs (Udry, 1996). ilarly, research in Ethiopia on the ability of farmers to identify the type and variety of sweet potato found that Recall periods for agricultural machinery and imple- 20 percent of farmers falsely reported local varieties ments usually focuses on the availability of assets over as improved, and 19 percent falsely reported improved the previous 12 months, differences in input use by varieties as local (Kosmowski et al., 2018). With respect crop-plot-season are important to capture, but may not to the quality of other inputs, Ashour et al. (2019) find be possible if the frequency of survey administration is that farmer’s perceptions of herbicide quality are over- annually, rather than seasonally. Machine age is usually stated, but that product quality is adulterated in their collected with the intention that depreciation might be sample. Fifteen percent of the herbicide samples they calculated, but much machinery depreciation depends on test are missing the active ingredient, while farmers use frequency and maintenance which are more difficult believe that 41 percent of herbicides are counterfeit in to capture in household surveys. 4. Agricultural Inputs 27 Table 5. Survey Decisions that inform input measurement POST-PLANTING QUESTIONNAIRE Module Key Design Choices Data quality Implications PP 5A - Household Unit of observation for labor This unit of observation makes possible the analysis of labor input Member Labor Inputs input is the individual-plot-level. at the plot-level and allows gender-disaggregation of labor input   Labor input data collected both in This reduces the length of the recall period for the respondents post-planting and in post-harvest thus reducing reporting error related to recall decay. visit. Respondent selection/recording In-person response considered more reliable than proxy- by individual-plot response, so respondent recording allowed to vary at plot- individual level. PP 5B - Hired and Plot-person type as unit of Makes possible analysis of labor inputs at the plot-level. Person Exchange Labor Inputs observation for hired and types are women, men, children under 15, whose labor inputs are exchange labor inputs considered to have varying returns. PP 7 - Seed Acquisition Crop as unit of observation for Allows analysis at plot-level for seeds which are key input in crop seed acquisition production. Reference period for seed The reference period aligns with the actual production cycle acquisition is ‘reference aiding respondent recall and improving data reliability. agricultural season’ Reporting of seed quantity in Reporting in standard units (e.g. kilograms) poses difficulties for non-standard units respondents and can worsen reporting error. Non-standard units should be country-specific. POST-HARVEST QUESTIONNAIRE Module Key Design Choices Data quality Implications PH 3 - Input Use Reference period for non-labor The reference period aligns with the actual production cycle inputs is ‘reference agricultural aiding respondent recall and improving data reliability. This season’ module is administered in the post-harvest visit because non- labor inputs may be used throughout the agricultural season. Unit of observation for non-labor Allows analysis at plot-level for non-labor inputs which are key inputs is the plot-level. input in crop production. Non-standard units of Reporting in standard units (e.g. kilograms) poses difficulties for measurement for reporting of respondents and can worsen reporting error. Non-standard units input quantities should be country-specific. Respondent selection/recording Decision maker may vary from plot to plot, data considered by plot most reliable if plot decision-maker can respond in person rather than proxy response. PH 4 - Input Roster Unit of observation for input Input is presumably acquired by type and not for each plot acquisition of non-labor inputs is separately making this unit of observation easier to recall. input type Reference period for non-labor The reference period aligns with the actual production cycle input acquisition is ‘reference aiding respondent recall and improving data reliability. This agricultural season’ module is administered in the post-harvest visit because non- labor inputs may be used throughout the agricultural season. PH 5A - Household Unit of observation for labor This unit of observation makes possible the analysis of labor Member Labor Inputs input is the individual-plot-level. input at the plot-level and allows gender-disaggregation of labor input Labor input data collected both in This reduces the length of the recall period for the respondents post-planting and in post-harvest thus reducing reporting error related to recall decay. visit. Respondent selection/recording Data are most reliable if each individual responds for him or by individual-plot herself personally, so respondent recording allowed to vary at plot-individual level. 28 AGRICULTURAL SURVEY DESIGN Module Key Design Choices Data quality Implications PH 5B - Hired and Plot-person type as unit of Makes possible analysis of labor inputs at the plot-level. Person Exchange Labor Inputs observation for hired and types are: women, men, children under 15, whose labor inputs exchange labor inputs are considered to have varying returns. PH 11A - Household Crop-individual as unit of Post-harvest labor is not usually tied to plot as the natural unit of Members Post Harvest observation production. Instead, post-harvest labor activities, such as shelling Labor or processing, may vary by crop. Respondent selection/recording Data are most reliable if each individual responds for him or by individual-crop herself personally, so respondent recording allowed to vary at crop-individual level. PH 11B - Hired/Exchange Crop-person type as unit of Makes possible analysis of labor inputs at the plot-level. Person Post Harvest Labor observation for hired and types are: women, men, children under 15, whose labor inputs exchange labor inputs are considered to have varying returns. PH 12 - Farm Past 12 months reference period Farm implements, machinery, structures in their lifetime are Implements, Machinery, independent of the agricultural season so that 12 months is a Structures more suitable length of reference period PH 13 - Extension Past 12 months reference period Extension services do not follow a seasonal pattern so that 12 Services months is a more suitable length of reference period AGRICULTURAL INPUTS: DESIGN types, days, hours per day and activities are included in FEATURES OF THE REFERENCE the post-planting labor modules. This survey design fea- QUESTIONNAIRES ture captures local context of reciprocal labor markets but makes calculating agricultural wages challenging due Input utilization is captured in both the post-planting to the sampling of households which may have low labor and post-harvest questionnaires. In the post-planting demand. Alternative labor market surveys or supplemen- questionnaire, labor used for land clearing and planting tal agricultural labor market modules may be essential to is recorded in module 5 as well as seed acquisition in accurately capture agricultural wage variation. module 7. In the post-harvest questionnaire, input use, labor inputs including household, hired and exchange Information on seed acquisition is found in module 7 of labor, farm capital, and extension service use is docu- the post-planting questionnaire. Seed acquisition informa- mented in modules 2-5 and 11-13. tion is linked to the crop codes, rather than alternative units of analysis such as the parcel-plot. As a survey design Labor in the post-planting questionnaire relates spe- choice, collecting seed information at the plot level may cifically to land preparation, planting, weeding, ridging be overly repetitive for the respondent, but the seed or fertilizing, and supervision labor. Labor inputs are acquisition module does allow for multiple seed types to recorded at the parcel-plot-individual level in the house- be recorded per crop. In this module, the source of seed hold labor input module 5A. The number of days and acquisition is the organizing principle of the module. Seed the average number of hours provided on the plot are can be left over from a previous harvest, acquired freely registered. In module 5B, hired and exchange labor is or purchased. recorded at the parcel-plot by person type (for instance men, women, and children under 15). The number of In the post-harvest questionnaire, input use related to person types used on the parcel-plot, the days, the hours production after planting is collected. Organic and inor- per day, activities and remuneration are documented. ganic fertilizer use, herbicide/pesticides use, and any With respect to remuneration, the value of payments equipment or machinery used for harvesting are popu- including those in-kind are asked. Exchange labor is also lated in the input roster collected at the parcel-plot level recorded at the parcel-plot level, but no estimation of in post-harvest questionnaire module 3. The input roster, remuneration is documented, only the number of person module 4, is organized by input type and documents 4. Agricultural Inputs 29 input purchases, quantities, and who financed inputs. Input The last input-related modules in the post-harvest acquisition such as being left over from the previous questionnaire include the agricultural capital and exten- season, own-produced inputs or free inputs are included sion services modules (13 and 14, respectively). The in accounting for quantities obtained by the household. agricultural capital module captures farm implements, machinery and structures that are inputs into agricultural Labor during the harvest period is documented in module production. The unit of analysis in this module is that of 5A for household labor and module 5B for hired and the item. The number, age, and sale value of the item is exchange labor. The questionnaire modules are similar to recorded. Any rental costs or rental revenues from such those of the post-planting labor module though post-har- items are also recorded in the questionnaire to facili- vest activities such as threshing/shelling and cleaning are tate cost analysis of production. The extension services expressly not recorded, so that this information can be module documents information received, who received collected within the post-harvest labor modules found information, and the frequency of contact with an exten- in modules 12A (household) and 12B (hired/exchange sive list of agricultural extension sources. These include labor). These post-harvest labor modules relate to agri- both formal sources such as government, NGO, private cultural labor activities such as threshing/shelling, drying, extension agents, or cooperatives, but also from peer cleaning, and processing. farmers and media sources. 30 AGRICULTURAL SURVEY DESIGN 5. Livestock Pastoral households rely on livestock assets and rev- counts, livestock indices such as tropical livestock units enue, but livestock are also central to rural and urban may be constructed to compare herds across households households in LMICs. While livestock is a vital compo- (FAO, 2011). It is important to note that the unit of anal- nent of the agricultural sector, data on livestock, costs ysis for most livestock modules in multi-topic household of production (veterinary services, labor, feed, and surveys is the household herd. Ownership of livestock shelter costs), income sources, mortality and livestock as with land modules has congruent measurement issues sales are often limited in nationally representative data. related to ownership and management of animals. As with Livestock modules are increasingly integrated into land modules, the livestock owner may not actually be the agricultural surveys to capture the multiple sources same person who manages the livestock. This presents of livestock use. In this chapter, we cover question- a potential opportunity for sex-disaggregated analysis, naire design choices in livestock modules highlighting linking to other sections of the multi-topic household measurement and empirical reasons for attention to questionnaire to infer effects on welfare if the owner and measurement error. Livestock measurement challenges livestock manager reside in the same household. relate to capturing the reproductive and production cycles of livestock systems, noting that the salience of In many pastoralist communities, sedentary households different farmer choices and continuous nature of pro- may own livestock that are managed by a nomadic duction create recall challenges in the context of annual herder or a local herder who manages several herds for multi-topic household surveys. Detailed livestock data households in a village. This raises multiple measurement require attention to both the stock and flow variables issues. First, the characteristics of the livestock manager in such a production system. Anagol et al. (2017) note in may not be available for analysis because the manager the context of India that median returns for cows were does not reside in the household and is not listed as -7 percent with a 17 percent median return for buffa- part of the household roster. Second, the manager may los when they assumed household labor was valued at derive compensation from animal products such as the zero. Despite mixed median returns, 51 percent and 45 milk of the animals. In a production function approach, percent of cow and buffalo owners reported negative revenues from animal sourced products should be listed returns. Measurement error may explain, in part, high as revenues from the livestock and payment in-kind to fractions of negative returns, underscoring the survey managers should be listed as a cost of production to design challenges of livestock measurement in house- the owners. In principle, an economic definition of pas- hold surveys. An LSMS guidebook on the livestock toral income would also include the net livestock value modules provides a more detailed review including change, but in the context of multi-topic household sur- variations in short, standard, and extended question- veys and due to the challenges of valuing livestock value naire modules (Zezza et al., 2016b). changes, we compare revenues and consumption to costs related to the ‘flow’ variables rather than changes MEASURING STOCKS in the value of the animal stock. If animals are herded remotely, then the unobservability of the use of animal A primary objective of livestock modules is to register products creates measurement issues. Third, changes in the number of animals that are owned by a household. animal stocks such as births, deaths or theft may also Description of the animal breeds, age, and sex are import- not be observable to the owner until herds are brought ant characteristics to quantify the household’s livestock closer to the sedentary household villages which usually portfolio. With the information on animal breeds and occurs seasonally. 5. Livestock 31 COSTS OF PRODUCTION would be considered as part of the household’s labor market activities, while activities related to producing As in much agricultural survey design, a production household public goods such as household cleaning or function approach to questionnaire design requires agri- cooking would be considered domestic activities, not cultural outputs to relate to inputs. Measuring livestock included in the labor market. Due to the interconnection inputs and the costs of livestock rearing are difficult of activities related to household activities and domestic for many of the same reasons that seasonal crop input animal care, women’s role in livestock production may be measurement is challenging. Some costs of livestock pro- undervalued in household surveys. duction are seasonal and many are irregular and involve high upfront costs. This is particularly true of veterinary Children’s work in livestock production both in terms costs which also illustrate how the unit of analysis can of herding and domestic production is likely under-re- matter in survey design. From a veterinary perspective, ported in agricultural surveys, though children play key immunizations and animal care costs relate to a specific roles in the livestock sector (FAO, 2013). While some animal’s health and are critical for different animals at dif- of children’s activities in the livestock sector may build ferent ages. In livestock modules collected for multi-topic their human capital, it may be considered child labor if household surveys, a household’s herd is recorded and the work includes hazardous activities or interferes with disaggregated by animal breed rather than at the animal their formal education. Formal measurement of children’s level.Veterinary services may be recorded either generally activities in livestock may be difficult in household-based for the herd or by breed. In practice, reporting veteri- surveys if child herders are not listed in the household nary services is often difficult for farmers due to longer roster or excluded due to household definitions that recall periods and the often specificity of veterinary ser- require minimum residency. vices required of individual animals rather than for an entire herd or breed. This has empirical implications on INCOME SOURCES whether an analyst can estimate a herd production func- tion or a livestock specific production function. Empirical The challenge of measuring revenue flows from stock objectives to measure either herd or livestock-specific variables requires carefully enumerating the multiple production functions need to be balanced against sources of revenue. For livestock, revenue flows from whether it is realistic for farmers to provide the neces- the sale of milk, eggs, animal traction and dung are all sary detail for livestock inputs and outputs at the chosen potential sources of revenue. Each is challenging as unit of analysis. Other costs of production such as food milk and egg production is irregular while demand costs are critical to understanding the return to livestock, for animal traction and dung are seasonal. The survey but many rural animals may graze freely while other ani- experiment on measuring milk off-take by Zezza et al. mals such as chickens or pigs may be confined to pens (2016a), notes that imputing milk off-take is difficult for their safety. In principle, the consumption of graz- because lactating females can be milked multiple times ing lands by the herd should be valued, but this analysis during the day, milking frequency varies seasonally, milk would require a land-based sampling frame rather than a production varies by lactation stage, and reproductive population-based sampling frame as in household surveys. and lactating females may not be milked. Milk produc- tion questions are often asked in a household survey The cost of labor devoted to animal care and herding by first soliciting the number of production months in is also critical to understanding the costs of livestock the last twelve months and then the average produc- production. Common issues with labor measurement, tion per month during production months. In addition reviewed in the earlier chapter, apply to livestock as well. to production-related measurement issues, valuing milk Time use is difficult to measure, particularly over recall production without milk sales information requires ana- periods longer than seven days as is attribution of labor lysts to make assumptions about sale frequency and milk time to certain animals. prices which may vary daily or weekly depending on the milk market. In the Zezza et al. (2016a) survey experi- For domestic animal production such as chickens, goats, ment, they find that 12-month recall using the standard or pigs, labor time devoted to animals may be inter- two-part household survey questions recorded similar spersed in the households’ daily activities. that includes production quantities relative to estimating a lactation production of household public goods. Chores such as curve which asks extended questions related to milk feeding, care and minor animal enclosure maintenance off-take at four different time periods during the annual 32 AGRICULTURAL SURVEY DESIGN recall period in an attempt to capture lactation dynamics. CHALLENGES TO LIVESTOCK Apart from this study, there are few survey experiments MEASUREMENT to inform the design of livestock modules in agricul- tural surveys. Despite measurement issues in milk sales, While Pica-Ciamarra et al. (2014) and Zezza et al. (2016b) estimates are used to inform important economic ques- review best practices in livestock module design, out- tions as in Casaburi and Macchiavello (2019) who use standing design issues are well documented in the survey contract design experiments in milk sales markets to design literature and remain challenges for agricultural estimate time preferences for Kenyan dairy farmers and survey design. Many of these challenges relate to the Hoddinott et al. (2015) who find a positive correlation household unit of analysis at which livestock data are in estimating the effects of cow ownership on children’s collected. First, mortality and livestock sales are often milk consumption and stunting. under-reported in household surveys as rural households often maintain herds outside of their residence. Mortality Seasonal sources of animal products such as animal trac- may be under-reported if stigma is associated with animal tion or dung can be captured in agricultural surveys if loss or animal owners are particularly sensitive to diffi- surveys are implemented during periods of high demand cult questions related to the death of animals. Livestock for such animal products. In practice, animal traction sales may not be frequent occurrences, so multiple sales services may be easier to value in active agricultural and at which prices may be difficult for respondents to services markets, most likely in contexts with contract recall. Second, as noted above, many livestock sector out- or irrigated agricultural schemes. Fully capturing dung comes should be recorded at the herd level rather than sales is difficult as larger animals are often encouraged to the household level, particularly for nomadic populations. freely roam in rural contexts to help fertilize the fields Household surveys with a population-based sampling and eat crop residues, clearing fields for farmers. As this will never capture herd level outcomes for which area- reciprocal arrangement between farmers and animal based sampling measures may be a better unit of analysis. owners is beneficial for both parties, formal dung mar- Lastly, FAO (2011) which outlines guidelines for preparing kets are rarer in rural contexts. Livestock holdings also livestock sector reviews underscores environmental and serve to provide insurance against risk by providing non- animal welfare issues that are often outside the scope crop sources of revenue from milk, egg or animal sales. of standard household livestock modules designed to Increased livestock holdings can improve food security measure a herd production function. If environmental through income diversification, but also providing a sav- consequences or social issues related to animal herding ings mechanism for rural households without access to are the objective of empirical analysis, area-based sampling formal financial services. may more accurately capture a representative sample of landscapes for which environmental consequences or conflict between pastoralists and farmers occur. Animal health and welfare are also not easily addressed when animal information is only captured at the animal breed level. Veterinary service expenditures may be capturing farmers engaged actively to help sick animals as opposed to evidence of animal welfare concerns. 5. Livestock 33 Table 6. Design Decisions in Livestock Production Measurement LIVESTOCK QUESTIONNAIRE LS SECTION A: Livestock breed as unit of Critical for accurate capture of herd size and composition Livestock Ownership observation and disaggregated information. by sex and age Valuation of livestock differs by gender and age. Greater aggregation would place burden on respondent to sum up various different breed. Differentiation between local/ Breed type has implication for valuation of livestock indigenous vs improved/exotic breeds Respondent recommended to Management responsibilities likely vary by groups of be recorded at the livestock livestock and so the most informed household member group (large ruminants, small may also vary at that level ruminants, poultry etc.) level Household member-level Recording who in the household owns livestock in key livestock ownership and to understanding gender dynamics and gaps in asset management responsibility ownership; additionally, ownership is differentiated from recorded management, which is also likely to have a gender gradient LS SECTION B1: Recall period for small and large Events related to large and small ruminants are likely not Changes in Stock over the ruminants: 12 months highly frequent, e.g. gestation periods for large ruminants is Past 12 Months 250-300 days and 140-160 days for small ruminants, making longer recall period best suited. LS SECTION B2: Recall period for poultry: 3 Events related to poultry are more frequent, e.g. gestation Changes in Stock over the months periods are only a few weeks, and there is limited Past 3 Months: Poultry seasonality, so that a shorter recall period is better suited. LS SECTION C: Unit of observation is livestock Management usually differs only between livestock groups Breeding, Housing, Water, group (large ruminant, small (large ruminants, small ruminants, poultry etc.) but not Feeding, and Hired Labor ruminants, poultry, etc.) between breed of the same group Recording of individual-level Important for gender analysis management responsibilities for different aspects of husbandry (breeding, water, housing, etc.) LS SECTION D: Unit of observation is livestock Health-related practices usually differ only between Animal Health group (large ruminant, small livestock groups (large ruminants, small ruminants, poultry ruminants, poultry, etc.) etc.) but not between breed of the same group LS SECTION E: Recall period 12 months Events related to large and small ruminants are likely not Milk Production (Off-take) highly frequent, e.g. gestation periods for large ruminants is 250-300 days and 140-160 days for small ruminants, making longer recall period best suited LS SECTION F: Recall period 3 months Clutching period is usually a few weeks and seasonality Egg Production limited so that shorter recall period better suited LS SECTION G: Unit of observation is livestock Limited variation between breeds within groups (large Animal Power group (large ruminant, small ruminants, small ruminants, poultry etc.). ruminants, poultry, etc.) LS SECTION H: Unit of observation is livestock Limited variation between breeds within groups (large Dung group (large ruminant, small ruminants, small ruminants, poultry etc.). ruminants, poultry, etc.) 34 AGRICULTURAL SURVEY DESIGN LIVESTOCK DATA: DESIGN FEATURES to have differential survival rates. Both modules pose OF THE REFERENCE QUESTIONNAIRES questions related to the birth, purchase, gifting of animals (both those received and given), lost, sold or slaughtered. The reference livestock questionnaire, which is sourced from Zezza et al. (2016b), is found in Appendix 3 which Animal care practices are covered in Section C of the also indicates a shorter base set of questions which are animal questionnaire. Animal care includes breeding prac- highlighted in green. For surveys which require compre- tices, housing, watering, feed and hired animal care labor. hensive data on livestock activities, the full questionnaire Recording animal care practices documents investments covers livestock ownership; changes in stocks (using a in animal quality, but also provides associated cost infor- 12-month-reference period for livestock and a three- mation which are essential in estimating an animal profit month reference period for poultry), breeding, housing function or the returns to livestock investment. Section water, feeding and hired labor; animal health; milk pro- D records animal health information specific to disease, duction; egg production; animal power; and dung. The curative care, and preventive practices such as vaccina- questionnaire modules, or sections as they are referred tion. This subsection records disease incidence as well as to in Zezza et al (2016b), and the related design decisions costs associated with vaccines and treatment. are summarized in Table 6. Sections E and F record animal-sourced revenues from The livestock ownership section of the questionnaire milk and egg production. Milk production is recorded by establishes which animals are owned by the house- animal where the average number of animals milked per hold, the number of owned animals, who within the month and the average amount of milk produced per day. household owns or manages the animals, and the main It is possible to record the person ID of the household reasons for holding animals. A key survey design choice member responsible for milking and recommended to is whether to limit questions to animal types and num- record the average amount and earnings from weekly milk bers within the household or to disaggregate livestock sold. The egg production module follows a similar struc- ownership and management responsibilities by person ture with deviations in the reference period. The eggs per ID within the household. Disaggregated livestock own- clutching period, the number of eggs sold and earnings ership and management questions permit analysis by over the past three months are recorded for poultry. sex, age and other individual characteristics and by animal sourced revenues. Sections G and H record in-kind animal uses which increase agricultural productivity via animal power and Sections B1 and B2 of the questionnaire classify the dung production. Animal power is recorded related to sources for changes in the stock of large and medi- both transportation and agricultural production. Earn- um-sized animals over a 12-month reference period and ings from animal services is recorded over the past 12 changes in the stock of poultry over a three-month ref- months. Animal dung uses are recorded for owned ani- erence period. Differences in reference periods reflect mals over the past 12 months as well. Animal dung uses assumptions concerning the reproductive cycles of include manure, fuel, construction material, feed to other animals, the liquidity of animal markets where smaller ani- animals, and sales. Any dung sales are recorded in local mals may be exchanged more frequently or more likely currency over the past 12 months. 35 6. Field Implementation Agricultural surveys may be implemented to achieve tradeoffs between data quality and implementation cost various objectives, and these objectives ought to be con- are present, include but are not limited to the following: sidered when designing and fielding data collection efforts. NSOs, which are common implementing agencies for agri- • Number of survey visits cultural surveys, may seek to use agricultural surveys to • Sample size produce estimates of crop production or monitor changes • Question translation in agricultural production over time. These objectives will • Integration of objective measures have implications on the survey design, as estimation of • Survey content and questionnaire design crop production may warrant a cross-sectional design • Mode of data collection with a larger sample in order to achieve estimates at the • Enumerator recruitment, training, and supervision desired level of representativeness, while an objective of monitoring and understanding the dynamics of the agricul- The reference questionnaire in Appendix 2, as well as tural sector warrants a panel approach. These objectives the discussion in previous chapters, promotes interviews and the implications on survey design all need to be con- being split up into two visits whereby households are sidered, while also balancing the often-binding resource visited once following planting and again following har- constraints and wider mandate of the agency. vest. This approach is recommended as it shortens the recall period for planting activities and splits the respon- This chapter discusses practical considerations for imple- dent burden across multiple visits. However, the feasibility menting agricultural surveys, with consideration for the of this approach depends on two things: (i) the agricul- primary objectives and constraints faced by implement- tural calendar of the given country, and (ii) the resource ing agencies, building on the experience of the LSMS-ISA. constraints and parallel survey operations of the imple- This chapter primarily focuses on implementation by menting agency which may further strain resource or in collaboration with NSOs, but many of the princi- availability. On the former, if a country has multiple crop- ples and challenges of collaboration are instructive for ping seasons per year, post-planting and post-harvest other national collaborations designed to produce data. visits may not be realistic for every cropping season, and For information on household surveys costs conducted depending on the agricultural calendar, recall periods may in partnership with NSOs around the globe, Kilic et al. be reduced in these multi-season years supporting a case (2017) summarizes costs and lessons learned from such for only a post-harvest visit for each season. Resource collaborations. constraints in terms of funding and/or human resources may also prohibit the implementation of a two-visit FIELDWORK IMPLEMENTATION survey structure, in which aggregation of data collection TRADEOFFS AND efforts into a single post-harvest visit would be the alter- SURVEY DESIGN CHOICES native. If resources allow, more frequent visits may be implemented to reduce recall bias. Survey fieldwork is a costly and complex operation. Here we discuss various design and implementation decisions The objectives of the survey will also inform the sample that are to be considered in order to achieve the survey size required. For operations aimed at highly precise pro- objectives while balancing data quality with resource duction estimates representative of small administrative constraints that are common to NSOs and other areas, a large sample size will likely be needed, depend- implementing agencies. These decision points, in which ing on the number of key crops and heterogeneity of the 36 AGRICULTURAL SURVEY DESIGN sample areas. For example, Ethiopia’s 2015-16 Annual ting). Where integration of these measurements on the Agricultural Sample Survey included a sample of more full sample is not possible due to resource constraints, than 44,000 households, while the LSMS-ISA supported a combination of sub-sampling for objective measure- Ethiopia Socioeconomic Survey panel in the same year ment and imputation techniques can be used to improve was implemented on a sample of roughly 5,000 house- data quality across the sample with limited financial holds.9 Tradeoffs will be necessary between the detail of investment. Kilic et al. (2017), for example, illustrate how the survey instrument and the sample size in order to multiple imputation techniques can be used to impute accommodate resource constraints of the implementing improved measures of plot area for observations where agency. The reference questionnaire provided in Appen- GPS measurement was not conducted. dix 2, for example, includes highly detailed information on the drivers of production that may not be feasible/real- A significant discussion with implementation partners istic for implementation in large sample sizes given the is the survey content and questionnaire design choices time it takes to administer such a survey. which may increase field time and interview complexity. Our discussion throughout this volume was to document There is also a tradeoff between carefully worded ques- best practice where possible and provide questionnaire tions and easily understandable questions. While survey designers with information about survey design choices. designers may be inclined to phrase questions in a very We fully acknowledge that it may not be possible to precise manner to get at the exact data they are seeking, implement best practice in each survey due to fieldwork this specific phrasing may either be difficult to under- constraints, budget or partner capacity, but these choices stand by respondents or lost in translation. Unless the should be taken with as much information on tradeoffs questionnaire is originally designed in the language in as possible. which it will be administered, translation will be neces- sary to account for the language, terms, and concepts of The reference questionnaire discussed and appended to the local context. Back translation, wherein the question- this document is designed to provide a detailed under- naire is translated to the local language, and then back to standing of agricultural activities. The thoroughness of the original language, is recommended to ensure proper this questionnaire, however, may not be necessary or translation. Piloting of the questionnaire will also serve to desired in contexts where agriculture makes up only flag incorrect translations or context-specific concepts a marginal share of the overall economy. In such cases that are not appropriately reflected in the questionnaire. where agriculture is not the focus of the survey or where it makes up a small share of household income, it is likely Significant tradeoffs exist around the measurement meth- still necessary to collect some data on agriculture, if only ods employed for specific topics. As discussed previously, to allow computation of total household incomes. The systematic bias is evident in the respondent reporting World Bank’s LSMS team has developed an agriculture of key agricultural variables such as land area and crop questionnaire which is designed for these cases, includ- production. These biases can be mitigated by the inte- ing only the minimum set of agricultural data necessary gration of objective measures into survey operations. for analyzing the role of farm activities in household These objective, or pseudo-objective, measures may livelihoods. This shortened agriculture questionnaire is include the use of GPS for area measurement, crop-cut- provided along with a reference household questionnaire ting for measurement of crop production, soil sampling and related guidebook (Oseni et al., 2020). for soil quality, etc. These measures have been shown to dramatically improve data quality, though they do come While the shortened version of the agriculture ques- at a cost. Costs include the procurement of relevant tionnaire includes crop, livestock, fishing, and forestry equipment, training on the measures, and additional field- production, it of course comes with tradeoffs in topical work time. The timing of visits should also be considered coverage and level of detail. Data on crop production when determining which objective measures to integrate is collected at the parcel level, rather than at the plot into a survey, as some are time sensitive and cannot be level, which shortens the questionnaire but eliminates completed in a single-visit survey (such as crop cut- the ability to conduct plot-level productivity analysis. By aggregating to the parcel level, the concept of the plot 9 manager, which is often used to conduct gender-related Sample size of the 2015-16 Annual Agricultural Survey extracted from the CSA’s Report on Area and Production of Major Crops, available analysis, is also omitted. Details on non-labor inputs are here. Sample size for the Ethiopia Socioeconomic Survey extracted from also significantly minimized, with binary questions asked the World Bank’s Microdata Catalog here. 6. Field Implementation 37 on fertilizer use rather than quantity of inputs applied, respondents and translated into an agricultural ques- for example. No information is collected on crop pro- tionnaire design. Question phrasing that better reflects cessing or extension services in the shorter module. cultural context will be easier for respondents to under- Additionally, the shortened questionnaire is designed to stand and accurately report their responses. be implemented in a single visit, in conjunction with the household survey, rather than split into post-planting and MODE OF DATA COLLECTION post-harvest visits. Agricultural survey data can be collected through The tradeoffs around mode of data collection and enumer- paper-based questionnaires (“paper assisted personal ators are discussed separately in the subsequent sections. interviewing” or PAPI) or tablet- or computer-based questionnaires (“computer assisted personal interview- PILOTING FOR IMPROVED ing” or CAPI). Historically, PAPI has been the method of SURVEY DESIGN choice, largely for lack of alternative options. However, CAPI technologies have evolved and are now widely Piloting surveys in the field before implementation is an adopted. At the time of writing, all LSMS-ISA surveys important tool to assess survey design choices. We high- have migrated from PAPI to CAPI implementation, using light the benefits of piloting particularly when a literature the World Bank’s Survey Solutions CAPI software. Both on a survey design choice is thin. Incorporating method- modes of data collection come with pros and cons, ological experiments into the pilot phase of a project is though CAPI implementation is strongly recommended. an approach to maximize the benefit of field research PAPI implementation offers the advantage of relatively for projects who have already budgeted for piloting. If less upfront preparation time prior to fieldwork but survey design effects are large enough to potentially bias requires significant effort in the data entry stage. In the estimates, low cost methodological experiments should case of PAPI implementation, it is recommended that a be able to detect survey design biases in relatively small double-data entry approach be used in order to limit sized experiments. the influence of entry errors. Data entry, which can be done concurrently in the field by a designated data entry While integrating methodological experiments into pilot- operator or at a centralized location, comes at a cost in ing provides a direct estimate of biases in survey design terms of personnel and time required for data dissemina- choices, piloting can take many forms with informative tion. Data quality controls in PAPI are also rather limited results for survey design. Piloting even with a few respon- relative to CAPI operations. In PAPI, supervisors often dents allows survey designers and national collaborators review the paper questionnaire at the close of each day. to assess a survey instrument’s ease of implementation If errors are found, this generally requires a re-visit to in the field, provide a method to evaluate interviewer’s the household in order to confirm or rectify the relevant understanding of the questionnaire, and identify areas responses. of questionnaire design improvement. Field protocols can be tested during pilots and a survey’s approximate CAPI-based data collection benefits from real-time duration can be estimated. During piloting, it is good to data quality controls as well as trackable revisions and evaluate survey duration using the pilot survey duration supervisor review. Research from various threads of as an upper bound on survey duration during the whole development economics have cited a reduction in errors of fieldwork as interviewers will become more efficient as a result of the use of CAPI implementation, for exam- with more questionnaire experience. ple the Asian Development Bank (2019) on labor force data and agricultural household survey data, Fafchamps Collaborations with national partners are enhanced by et al. (2012) on microenterprise profit data, and Caeyers integrating survey designers with national stakeholders et al. (2012) on consumption data. Errors that are caught in the piloting process. National experts in complemen- in real-time, through warning messages for pre-coded tary disciplines can also provide insights to better design errors, prompt the enumerator to confirm or correct surveys tailored to respondent thinking about variables relevant inputs at the moment the question is asked, of interest for researchers. Field partner collaboration thereby increasing data quality and limiting the need to extends beyond soliciting accurate response categories revisit the household. Data quality can be increased at for questions and into how variations in production sys- the micro level, but also at the macro level through var- tems within countries are best represented clearly for ious survey management tools that allow office-based 38 AGRICULTURAL SURVEY DESIGN personnel to effectively monitor operations and data field supervision and continuous data review. Where a quality in real time. With relatively minimal investment, panel approach is employed, a specialized tracking team survey management tables and graphics can also be can be deployed with the objective of tracking and inter- automated through application programming interfaces viewing respondents that have relocated between survey (APIs). An additional advantage of CAPI is the possibil- rounds and/or split off from previous households and ity of developing and administering interviews in different created new split-off households.10 In designing fieldwork languages, allowing interviewers to toggle between differ- structure, enumerators may be mobile (moving with teams ent languages as necessary to accommodate respondents. from enumeration area to enumeration area) or resident Implementation via CAPI requires additional preparation (permanently residing in or close to the enumeration area time in order to test the application and program error for which s/he is responsible). The survey-structure and checks, but it can significantly decrease the time from the objectives of the survey should help to inform the field- close of fieldwork to data dissemination as data entry is work structure. Where multiple visits are necessary, such not needed and data quality controls can be implemented as when crop-cutting is to be implemented, resident enu- continuously throughout fieldwork. Numerous open- merators may be more cost-efficient and effective. source and private CAPI software options exist, common platforms include the World Bank’s Survey Solutions At the base of the data generating process is the enu- software, SurveyCTO, Open Data Kit (ODK), CSPro and merator who manages the collection of data from SurveyBee. The questionnaires found in Appendix 2 and 3 respondents. A meta-analysis of the literature11 establishes are also publicly available in Survey Solutions. that enumerator behavioral traits and demographic char- acteristics influence survey responses and by extension With the rate of mobile phone ownership growing glob- data quality. Response rates and response biases are par- ally, the potential for phone-based or phone-assisted ticularly influenced by specific enumerator characteristics surveys increases. While national-scale agricultural data (age, sex, ethnicity, experience, education, etc.), behaviors collection through phone surveys has yet to be imple- (formal versus conversational interview style), cognitive mented, recent methodological research on the topic and non-cognitive skills (such as mathematical ability, has shown promise for improving data quality in some reading, attention to detail, and empathy) and interviewer areas, while the flurry of phone surveys implemented in experience12,13,14. Recent research on enumerator effects the face of the COVID-19 pandemic shows promise for by Di Maio and Fiala (2019), through a randomized exper- feasibility of implementation. Methodological research on iment in Uganda, illustrate that enumerator characteristics agricultural labor data collection in Tanzania, for example, and their differences from respondent characteristics shows consistent data collected through weekly in-per- affect survey responses and ultimately data quality, espe- son visits, considered to be the gold standard, and weekly cially related to sensitive political topics. Responses in phone calls, suggesting that phone surveys could be a rea- less sensitive topics were much less, or not at all, sensi- sonable alternative (Arthi et al., 2018). Implementation of tive to enumerator characteristics. This is supported by surveys via mobile phone will necessitate simplification of the questionnaire instrument and careful assessment of mobile phone coverage in the population of interest. Fur- 10 Tracking protocols for the LSMS-ISA surveys are available as part of the documentation released with the survey data. Visit the World Bank’s ther research is needed in order to fully understand the Microdata Library (https://microdata.worldbank.org/ ) to find all of risks and challenges of phone-based national-level agricul- the LSMS-ISA surveys. The tracking forms and protocols (included in tural survey data collection. the enumerator’s manual) for the Malawi Integrated Household Panel Survey, for example, are available here: https://microdata.worldbank.org/ index.php/catalog/2939/related-materials FIELDWORK ORGANIZATION 11 West, B.T. and Blom, A.G., 2017. Explaining interviewer effects: A research AND LOGISTICS synthesis. Journal of Survey Statistics and Methodology, 5(2), pp.175-211. 12 Cannell, C. F., and Laurent. (1977). A summary of studies of interviewing Several features of fieldwork organization and logistics methodology. Vital and Health Statistics: Series 2, Data Evaluation and Methods Research (No. 69. DHEW Publication No. (HRA), 77-1343). can impact data quality outcomes, including the composi- Washington, DC: U.S. Government Printing Office. tion, training, and structure of fieldwork teams. Fieldwork 13 Fowler,F. J. Jr., and Mangione, T. W. (1985). The value of interviewer organization should include a supervisory hierarchy, training and supervision (Final report to the National Center for Health services Research). Boston, MA: Center for Survey Research. with several field teams deployed, each including a set of 14 Alcser K., Clemens J., Holland L., Guyer H., Hu M. 2016. Interviewer enumerators and a supervisor, and a separate headquar- recruitment, selection, and training. Cross-Cultural Survey Guidelines. ters-based supervisory team which conducts periodic University of Michigan. 6. Field Implementation 39 additional research which suggests that salience and sen- tion and concepts are clear and harmonized prior to the sitivity of the questions at hand influence the nature and full training. Training of trainers, and potentially fieldwork magnitude of enumerator effects (for example, Himelein, trainings, could also be held virtually in-part, if the con- 2016; Laajaj and Macours, 2017). Beaman et al. (2018) find text allows. referral-based recruitment may disadvantage women in an enumerator recruitment experiment in Malawi. Marx et al. Fieldwork training should be inclusive of all facets of (2018) provide evidence on team composition and ethnic fieldwork, from protocols on conduct with respondents diversity on enumerator performance. Data on time use to questionnaire content to data review, revision, and of field teams suggests horizontally homogenous teams submission. This generally requires both classroom-style (enumerator teams) organized tasks more efficiently, training on relevant concepts, questionnaire content while vertically homogenous teams (supervisors, enumer- and layout, and CAPI software, if applicable. Classroom ators, data monitors) had lower effort. Additional research instruction should be complemented by mock inter- is needed to gauge the sensitivity of agricultural data spe- views, both amongst the enumerators themselves and cifically to enumerator characteristics. with pilot households, and training on objective measures that are being integrated into the survey. For example, The capacity of enumerators should also be carefully training on the use of GPS devices for area measurement assessed to ensure they possess the skills needed for requires theoretical training in the classroom followed the specific survey operation, with an eye for the data by repeated practice outside the classroom.15 While the collection mode being employed. Some literature exists integration of objective measures is intended to reduce to suggest that enumerator recruitment and contract subjectivity in responses, training is still required to real- structure affect data quality and enumerator job perfor- ize their maximum benefit. mance. For agricultural surveys, knowledge of farming may prove to be an asset in terms of both understanding Finally, enumerator activities should continue to be mon- the questionnaire content and establishing rapport with itored throughout fieldwork. Supervision is particularly respondents. This should be weighed against enumerator important at the onset of fieldwork, to identify and cor- education, language skills, and familiarity with technologi- rect any misunderstandings, but is relevant for the full cal tools (in the case of CAPI implementation). duration. Incoming data should be monitored both by supervisors in the field and the survey management team Effective survey implementation relies on a homogenous, at the headquarters level, made quicker and more auto- thorough understanding of the questionnaire content and mated through the use of various CAPI tools. Additionally, concepts. Training of field teams, as well as those involved supervisors from both the field and headquarters should in survey management at the headquarters level, is essen- conduct random spot checks or call backs to households tial in developing this thorough understanding. Depending to confirm data entered by enumerators matches that on the number of people involved and the structure of reported by the household. the NSO or other implementing agency, training can be conducted in a single centralized location or in several PLOT VISITS AND GEOREFERENCING parallel decentralized training locations. Centralized trainings benefit from consistent messaging for all train- An additional layer of consideration for the design and ees, though may be ineffective if there is a large number implementation of an agricultural survey is whether vis- of trainees. Decentralized trainings are often necessi- itation to agricultural plots (or parcels) will be carried tated by large groups of trainees and/or decentralized out. Visitation of agricultural plots, whereby an enumer- statistical offices, though they come with the complica- ator physically travels to plots typically with a member tion of ensuring discussions and questionnaire revisions of the sampled household or farm, enables a wide range are communicated across locations. Irrespective of the of additional survey operations and objective measures. number of training locations, a training should be held Crop-cutting, which has been shown to significantly prior to each survey visit (one training for the post-plant- reduce measurement error in agricultural production esti- ing visit and one for the post-harvest visit, in the case mates, as discussed in Chapter 3, for example, requires at of a two-visit survey). Prior to each training, a “training of trainers” is commonly held, especially in the case of 15 Forguidance on the implementation of GPS for area measurement, see decentralized trainings, in order to train the facilitators the LSMS guidebook on Land Area Measurement in Household Surveys of the field team trainings and ensure the implementa- (Carletto et al., 2016). 40 AGRICULTURAL SURVEY DESIGN least two visits to the selected plots (after planting and at brated model of maize yields. Related research by Lobell time of harvest). Objective measures of agricultural land et al. (2020) highlight the possibility of sub-sampling of area, also discussed in Chapter 3, require physical pres- crop-cutting-based production measurement, whereby ence on the land. only a strategically selected subsample of plots are sub- ject to crop-cutting rather than the full sample, could be In addition to the benefits of potential objective mea- sufficient to calibrate satellite-based estimation models. surements, visitation of agricultural plots allows for the This would greatly improve the feasibility of scale-up in georeferencing of plots (the collection of GPS coordinates national survey operations. of the plots and/or the identification of the boundaries of the plots), which can add significant value through the While georeferencing of agricultural plots is strongly integration with other geospatial and remotely-sensed recommended when feasible, considerations need to be data sources such as data on soil properties, forest cov- made to protect the confidentiality of survey respon- erage, climate, and water quality, among many others. dents. GPS coordinates should not be disseminated to Integration of these properties with data on farming the public without an anonymization procedure.16 Ano- practices and outcomes can provide more complete nymization, or offsetting, of GPS coordinates will reduce insights into the agricultural landscape and barriers to the precision of the coordinates and therefore limit the increased productivity. value of integrated data sources through loss of variation at the local level. With raw GPS coordinates one is able Integration of agricultural survey data with various types to extract climate or soil conditions for that particular of spatial and Earth observation data, possible with point, for example, whereas offset coordinates are com- the georeferencing of plots, has expanded the scope monly provided at the enumeration area level and thus of analysis possible with the use of survey data, while the data extracted from geospatial data will be the same also improving the performance prediction algorithms across the enumeration area. To maximize the value of employed in spatial analyses. In recent years there has integration with geospatial data sources while preserv- been growing demand for crop yield estimation via ing respondent confidentiality, the LSMS-ISA prepares remotely sensed imagery, for example, for more fre- and disseminates a set of geovariables that are derived quent and rapid assessment of production estimates from the raw GPS coordinates, but does not release when agricultural surveys or censuses are not available. the raw coordinates themselves. Additional research Recent research has cited improvements in yield esti- is ongoing with respect to further ways of maximizing mation when estimation algorithms are calibrated on the use of georeferenced data while maintaining confi- ground-based crop-cutting data collected though agri- dentiality. An active literature on creating synthetic data cultural surveys. Lobell et al. (2019) find that calibration from confidential data may provide some insights to based on crop-cutting data from Uganda significantly data integration challenges (see for example, Abowd and limited the overestimation bias observed in the uncali- Schmutte 2019). 16 An example of a GPS anonymization protocol employed by the DHS program can be found here. 41 7. Conclusions and Future Directions for Agricultural Survey Design In the 20 years that followed the Ghosh and Glewwe of household survey and alternative data sources, such as (2000), Designing Household Survey Questionnaires for telephone surveys, administrative records and geospatial Developing Countries volumes, innovations in agricul- data, can result in analytical gains from both sources and tural survey design have expanded the set of answerable more precise measurement of variables when self-re- policy and research questions by filling data gaps, along ports are not reliable. Household survey data, therefore, with improving the quality of descriptive statistics that should be viewed as a complement to other data are critical to development monitoring and planning. This sources, not a substitute. Integrating supplementary data volume has focused on an approach to agricultural survey with household surveys is a trend for which increased design that considers not only the objective of limiting investment by NSOs, researchers, international organi- measurement error through survey design, but also the zations and policymakers would yield a wider range of tradeoffs that NSOs, researchers, policymakers and other policy analysis. The integration of remotely sensed agri- stakeholders face when designing and using agricultural cultural, ecological, and climate data is not uncommon data. While the recommendations and design choices in agricultural analysis, but the costs of such integration found in the reference questionnaires have evolved from are often high. In Chapter 3, we highlight the potential of the LSMS team’s collaborations with NSOs, different stud- remotely sensed data. Here we also highlight limitations ies and partnerships may find different data requirements in terms of validation study requirements to calibrate are useful for achieving alternative survey objectives. Agri- remotely sensed data for ‘ground-truthing’ and the lim- cultural survey design must consider the real tradeoffs itations of what can be imputed from remotely sensed that field implementation and partnerships necessitate crop images, particularly for tree and tuber crops and in concert rather than independently of the economet- in intercropped areas. While these limitations are not ric consideration of minimizing measurement error to trivial, remotely sensed plot data would permit a much increase statistical precision and improve internal validity. leaner agricultural survey instrument which may improve self-reported variables on land use and production infor- In this chapter, we take stock of future directions in agri- mation. Without household survey integration, remotely cultural survey design which may help guide a future sensed data itself does not satisfy most data require- methodological research agenda as well as stimulate ments for policy analysis which require linking inputs and debate about best practices. We highlight opportunities the use of production. to expand the application of publicly collected data sets through data integration and the challenges of measuring With careful design, agricultural survey data also has the fundamental agricultural concepts such as farmer prefer- potential to be integrated with administrative or census ences, labor, profitability and land. data. With the proper identifiers in the survey data to allow matching between the two sources, survey data can DATA INTEGRATION be merged, for example, with firm data linking employee- and employer-reported data, such as that done by de The value of agricultural survey data can be increased Nicola and Giné (2014) for fishing communities in India, through integration with other data sources. Integration or with microfinance or agricultural lender data to match 42 AGRICULTURAL SURVEY DESIGN access to credit with agricultural practices and outcomes. selection biases, particularly in the context of agricultural Agricultural administrative data are often under-utilized, surveys, need to be further explored though improved in part, because extension data systems may not be well telephone survey methodologies offer the potential for developed or data systems in Ministries of Agriculture high frequency agricultural data collection with significant may not be always linked to the National Statistical Office. cost savings. Unit of analysis and proxy response are key Particularly in irrigated agricultural areas or commercial open concerns for telephone surveys that use random agricultural settings, administrative data on inputs such as digit dialing. water use, technology adoption, or crop yields can be of high quality. Where household surveys may also have the With integration of supplemental data sources, how- most promise in integrating with extension data sources ever, comes increased concern of respondent anonymity. are on topics of agricultural and environmental practices. While open data is the mantra of development institu- A promising opportunity for integration may come with tions at large, as well as NSOs, respondent confidentiality the expansion of digital, mobile and smart phone-based must be maintained. This entails removing all informa- delivery of agricultural extension services, discussed for tion that could lead to the identification of respondents, instance in Fabregas et al. (2019). including names, phone numbers, ID numbers, etc. How- ever, integration with additional data sources complicates A promising innovation with potential applications for the anonymization practice, as it becomes easier to measuring physical activity, productivity and time use is re-identify individuals through triangulation, especially wearable technology such as accelerometers worn either those with uncommon characteristics (Heffetz and Ligett, specifically for measurement purposes or embedded 2014). Maximizing the value of survey data through public in mobile phones. Akogun et al. (2020) validate physical dissemination and integration with other sources, while activity measures relative to individual farmer harvesting maintaining the privacy of respondents, will encourage productivity, demonstrating that physical activity mea- sustainable and effective survey operations going forward. sures are highly correlated with work effort. They also document potential uses and limitations of accelerom- MEASURING THEORETICAL eters for physical activity and productivity measures. CONCEPTS MORE PRECISELY Friedman et al. (forthcoming) illustrate a use case for physical activity data in understanding intra-household The Ghosh and Glewwe volumes highlighted innovations energy expenditure in rural settings. As the measure- in economic theory and how these ideas were translated ment of labor variables is one of the most challenging to survey questions, generating the data for empirical self-reported survey design problems, improved data analysis. As we have highlighted earlier, innovations in integration using accelerometers and household surveys relating consumption measures to welfare (Deaton and remains promising. Zaidi, 2002), labor supply, returns to education, and the agricultural production function, among others, have Lastly, we highlight telephone surveys as a survey mode resulted in improved survey design. There is still more that could be integrated to measure higher frequency work to do as an evolving scientific field to improve outcomes such as labor, agricultural sales or food survey design to facilitate empirical analysis. We highlight stocks. Telephone surveys have been heavily integrated a few areas of promise below but underscore that these into field operations in response to the global COVID- examples are not an exhaustive research agenda. 19 pandemic which resulted in limitations on in-person interviews. Literature on telephone surveys in developing Longstanding difficulties in measuring agricultural time countries existed before the pandemic (see for example, use on plots or with livestock, not only create data that Dabalen et al. 2016), but the global pandemic has resulted may be biased by non-classical measurement error but in a renewed interest in telephone survey methodology. have implicitly restricted the set of potential econometric Existing literature suggests that response rates vary sub- applications. The profit function is an important example stantially by mode. Some response rates are 250 percent of how restrictions in data quality prevent estimation of higher in computer-assisted telephone interviews (CATI) an underlying theoretical concept. Agricultural survey than interactive voice response surveys or SMS surveys design has often adopted a production function approach based on metadata from 41 studies in 20 countries, but as input quantities and outputs are more easily recalled are significantly improved with small incentives (IPA, by respondents. The limitation of such an approach, par- 2020). Drawbacks related to response rates and potential ticularly for the LSMS-ISA surveys founded to measure 7. Conclusions and Future Directions for Agricultural Survey Design 43 welfare and its determinants, is that farmers do not max- OPEN MEASUREMENT QUESTIONS imize production and production does not necessarily FOR VALIDATION RESEARCH relate to higher welfare as demonstrated in the food security and technology adoption literatures. The policy Our review of the agricultural survey design literature implications of not measuring profitability are potentially leaves us impressed by the substantial progress the large for agricultural development. research, policy and national statistical office stakehold- ers have achieved since the Ghosh and Glewwe (2000) As highlighted in Chapter 3, markets are often difficult to volume on measuring agricultural information. The range capture in household surveys as households only repre- of empirical applications that innovations in plot and sent a partial segment of the demand and supply present sex-disaggregated data in agricultural survey design have in an economy. Recording the relevant prices in input, opened is impressive.We highlight that agricultural survey output, farm-gate, and market prices are vague boundar- design is in no way a closed topic. A large and important ies in a continuum of economic activity and overlapping research agenda of open measurement questions remains markets. The difficulty of accurately measuring prices and with important research and policy implications for farm- wages often limits potential structural estimation analy- ers throughout the world. sis. The concept ‘market access’ is inherently flawed when markets exist but may be seasonal or thin. For trans- In different sections of this volume, we have highlighted actions that are infrequent such as in land or housing open measurement questions. Agricultural survey design markets, the most valuable household assets may not be is predicated on land measurement. Though a large lit- well captured in household survey samples but be the erature has focused on the biases of self-reports, open most consequential transactions from a welfare perspec- questions on remotely sensed land measures, including tive. High frequency transactions are also consequential the effect of slope and the feasibility of measurement at to household welfare but may be difficult to measure. the plot level in smallholder farming systems, are import- Examples of this include food consumed outside of the ant to further improve this literature. From a production household which may be difficult to record in household perspective, open production measurement questions surveys by a main respondent who does not observe related to root, tuber and tree crops as well as appropri- food consumption outside of the household (Oseni ately attributing intercropped plots for yield estimation et al., 2017). In the case of small-scale food producers, are important production systems with remaining mea- frequent small transactions may be difficult to recall for surement challenges. With respect to agricultural output, respondents, particularly as recall periods increase. Food post-harvest loss remains an active area of measure- markets are fundamental to the agricultural supply chain, ment research, particularly as it relates not only to food but their measurement remains challenging as repre- waste, but improved profitability measures. In the live- sented in the above examples. stock sector, livestock labor measurement is even more challenging due to unit of analysis and attribution to par- As a last example, we cite the measurement of prefer- ticular livestock in both sedentary and pastoralist systems ences and agency as two theoretical concepts that have of production. There continues to be much to learn with been widespread importance in empirical analysis. While respect to alternative survey modes and the integration the literature on risk preferences is extensive in the of survey modes, particularly when face-to-face inter- laboratory setting, preferences, including risk and time views may be limited. As stakeholders have responded to inconsistency which are foundational to understanding public demands to make nationally representative data poverty and investment dynamics, remain omitted in publicly available, a concurrent set of privacy and geo-ref- most nationally representative survey data. The insights erencing challenges are important to balance against the and econometric applications from behavioral economics empirical possibilities facilitated by such data. We under- have not been fully integrated into multi-topic household score that earlier measurement challenges, such as those questionnaires. A more widely emphasized, but not fully that were identified after the Ghosh and Glewwe (2000) adapted literature on women’s empowerment (Alkire et volumes have led to innovations in agricultural survey al., 2013; Glennester et al., 2018) is an important example design. The continued interaction of not a sole set of of the importance of improved measures of agency to stakeholders, but the collaboration between research, complement sex-disaggregated data that is largely avail- policy, national statistical office, and donors will facilitate able in nationally representative surveys. future innovations in agricultural survey design. 44 AGRICULTURAL SURVEY DESIGN References Abay, K.A., Abate, G.T., Barrett, C.B., and Bernard, T. 2019. “Correlated Non-Classical Measurement Errors, ‘Second Best’ Policy Inference and the Inverse Size-Productivity Relationship in Agriculture.” Journal of Development Economics 139(C), 171–184. Akogun, Oladele, Andrew Dillon, Jed Friedman, Ashesh Prasann, and Pieter Serneels. 2021. “Productivity and Health: Physical Activity as a Measure of Effort,” World Bank Economic Review, forthcoming. Alkire, S., Meinzen-Dick, R., Peterman, A., Quisumbing, A., Seymour, G., and Vaz, A. 2013. “The Women’s Empowerment in Agriculture Index.” World Development 52, 71–91. Ambler, K., Doss, C., Kieran, C., and Passarelli, S. 2017. “He says, she says: Exploring Patterns of Spousal Agreement in Bangladesh.” International Food Policy Research Institute. Anagol, S., Etang, A., and Karlan, D. 2017. “Continued Existence of Cows Disproves Central Tenets of Capitalism?” Economic Development and Cultural Change 65(4), 583-618. Antman, F., and McKenzie, D. J. 2007. “Earnings Mobility and Measurement Error: A Pseudo‐Panel Approach.” Economic Development and Cultural Change 56, 125-161. https://doi:10.1086/520561 Arthi,V., Beegle, K., De Weerdt, J., and Palacios-López, A. 2018. “Not Your Average Job: Measuring Farm Labor in Tanzania.” Journal of Development Economics 130, 160-172. Asian Development Bank. 2019. “The CAPI Effect: Boosting Survey Data through Mobile Technology. A Special Supplement of Key Indicators for Asia and the Pacific 2019.” Manila: ADB. Ashour, M., Gilligan, D.O., Hoel, J.B., and Karachiwalla, N.I. 2019. “Do Beliefs About Herbicide Quality Correspond with Actual Quality in Local Markets? Evidence from Uganda.” Journal of Development Studies 55, 1285–1306. https://doi. org/10.1080/00220388.2018.1464143 Bardasi, E., Beegle, K., Dillon, A., and Serneels, P. 2011. “Do Labor Statistics Depend on How and to Whom the Questions are Asked? Results from a Survey Experiment in Tanzania.” World Bank Economic Review 25, 418–447. Bautista, R. 2012. “An Overlooked Approach in Survey Research: Total Survey Error.” In Handbook of survey methodology for the social sciences (p. 37–49). Springer. Beaman, L. and Dillon, A. 2012. “Do Household Definitions Matter in Survey Design? Results from a Randomized Survey Experiment in Mali.” Journal of Development Economics 98, 124–135. https://doi.org/10.1016/j.jdeveco.2011.06.005 Beegle, K., Carletto, C., and Himelein, K. 2012. “Reliability of Recall in Agricultural Data.” Journal of Development Economics 98, 34–41. Benedetti, L., De Baets, B., Nopens, I., and Vanrolleghem, P. A. 2010. “Multi-criteria analysis of wastewater treatment plant design and control scenarios under uncertainty.” Environmental Modelling and Software 25, 616-621. https:// doi:10.1016/j.envsoft.2009.06.003 Benes, Elisa M., and Kieran Walsh. 2018a. “ILO LFS Pilot Studies in Follow Up to the 19th ICLS: Background, Objectives, and Methodology.” Statistical Methodology Series 1 (May). Geneva; International Labour Office. https://www.ilo.org/ wcmsp5/groups/public/---dgreports/---stat/documents/publication/wcms_627873.pdf. Benes, Elisa M., and Kieran Walsh. 2018b. “Measuring Employment in Labour Force Surveys: Main Findings from the ILO LFS Pilot Studies.” Statistical Methodology Series 4 (July). Geneva; International Labour Office. https://www.ilo.org/ wcmsp5/groups/public/---dgreports/---stat/documents/publication/wcms_635732.pdf. References 45 Berazneva, J., McBride, L., Sheahan, M., and Güereña, D. 2018. “Empirical Assessment of Subjective and Objective Soil Fertility Metrics in East Africa: Implications for Researchers and Policy Makers.” World Development 105, 367-382. Bevis, L.EM., and Barrett, C.B., 2020. “Close to the Edge: High Productivity at Plot Peripheries and the Inverse Size- Productivity Relationship.” Journal of Development Economics 143, 102377. Bowley, A. L., and Burnett-Hurst, A. R. (1926). Livelihood and Poverty (London, 1915). AL Bowley and Margaret Hogg, Has Poverty Diminished, 15. Brown, J. F., and Pervez, M. S. 2014. “Merging Remote Sensing Data and National Agricultural Statistics to Model Change in Irrigated Agriculture.” Agricultural Systems 127, 28-40. Caeyers, B., Chalmers, N., and De Weerdt, J. 2012. “Improving Consumption Measurement and Other Survey Data through CAPI: Evidence from a Randomized Experiment.” Journal of Development Economics 98, 19–33. Carfagna, E., and Gallego, F. J. (2005). “Using Remote Sensing for Agricultural Statistics.” International statistical review 73(3), 389-404. Carletto, C., Aynekulu, E., Gourlay, S., and Shepherd, K. 2017a. Collecting the Dirt on Soils: Advancements in Plot-level Soil Testing and Implications for Agricultural Statistics. World Bank Policy Research Working Paper Series No. 8057. World Bank, Washington, DC. Carletto, C., Gourlay, S., Murray, S., and Zezza, A. 2017. “Cheaper, Faster, and More than Good Enough: Is GPS the new gold standard in land area measurement?” Survey Research Methods (v11) No. 3: 235-265. Carletto, C., Gourlay, S., and Winters, P. 2015. “From Guesstimates to GPStimates: Land Area Measurement and Implications for Agricultural Analysis.” Journal of African Economics 24, 593–628. Carletto, G., Himelein, K., Kilic,T., Murray, S., Oseni, M., Scott, K. and Steele, D. 2010. “Improving the Availability, Quality and Policy-Relevance of Agricultural Data: The Living Standards Measurement Study–Integrated Surveys on Agriculture.” In Third Wye City Group Global Conference on Agricultural and Rural Household Statistic. Washington, DC. Carletto, G., Savastano, S., and Zezza, A. 2013. “Fact or Artefact: The Impact of Measurement Errors on the Farm Size - Productivity Relationship.” Journal of Development Economics 103, 254–261. https://doi.org/10.1016/j. jdeveco.2013.03.004 Carletto, G., Gourlay, S., Murray, S., and Zezza,A. 2016.“Land Area Measurement in Household Surveys: Empirical Evidence and Practical Guidance for Effective Data Collection (English).” LSMS Guidebook Washington, D.C.; World Bank Group. http://documents.worldbank.org/curated/en/606691587036985925/Land-Area-Measurement-in-Household- Surveys-Empirical-Evidence-and-Practical-Guidance-for-Effective-Data-Collection Casaburi, L., and Macchiavello, R. 2019. “Demand and Supply of Infrequent Payments as a Commitment Device: Evidence from Kenya.” American Economic Review 109, 523-55. https://doi:10.1257/aer.20180281 Chambers, R. G. 1988. Applied Production Economics: a Dual Approach. Cambridge University Press. Chambers, R. G., and Quiggin, J. 2000. Uncertainty, Production, Choice, and Agency: the State-Contingent Approach. Cambridge University Press. Coates, J.C., Webb, P., Houser, R.F., Rogers, B.L., and Wilde, P. 2010. “He Said, She Said: Who Should Speak for Households about Experiences of Food Insecurity in Bangladesh?” Food Security 2, 81–95. Converse, J. M., and Presser, S. 1986. Quantitative applications in the social sciences. Survey Questions: Handcrafting the Standardized Questionnaire. Newbury Park, CA: Sage Publications, 10. Dang, H. L., Li, E., Nuberg, I., and Bruwer, J. 2014. Understanding Farmers’ Adaptation Intention to Climate Change: A Structural Equation Modelling Study in the Mekong Delta, Vietnam. Environmental Science and Policy 41, 11-22. https:// doi.org/10.1016/j.envsci.2014.04.002 De Groote, H. and Traoré, O. 2005. “The Cost of Accuracy in Crop Area Estimation.” Agricultural Systems. 84, 21–38. De Heer,W., de Leeuw, E.D., and van der Zouwen, J. 1999. Methodological Issues in Survey Research: a Historical Review.” Bulletin of Sociological Methodology/Bulletin de Méthodologie Sociologique 64, 25–48. De Nicola, F., and Giné, X. 2014. “How Accurate are Recall Rata? Evidence from Coastal India.” Journal of Development Economics 106, 52–65. 46 AGRICULTURAL SURVEY DESIGN De Weerdt, J., Beegle, K., and Gibson, J. 2019. “What Can We Learn from Experimenting with Survey Methods?” (No. 41819), LICOS Discussion Papers. LICOS - Centre for Institutions and Economic Performance, KU Leuven. Deaton, A. 1985. “Panel Data from Time Series of Cross-Sections.” Journal of Econometrics 30, 109-126. https://doi. org/10.1016/0304-4076(85)90134-4 Deaton, A. and Zaidi, S. 2002. Guidelines for constructing consumption aggregates for welfare analysis. Washington, DC: World Bank. Deininger, K., Carletto, C., Savastano, S., and Muwonge, J. 2012. “Can diaries help in improving agricultural production statistics? Evidence from Uganda.” Journal of Development Economics 98, 42–50. Desiere, S. 2015. “Area Measurement in Agricultural Surveys: GPS or Compass and Rope?” In S. Desiere (Ed.), “From Raw Numbers to Robust Evidence. Finding Fact, Avoiding Fiction” (Chapter, Area Measurement in Agricultural Surveys: GPS or Compass and Rope?). Doctoral dissertation, Ghent University. Desiere, S., Jolliffe, D. 2018. “Land Productivity and Plot Size: Is Measurement Error Driving the Inverse Relationship?” Journal of Development Economics 130, 84–98. Di Maio, M., and Fiala, N. 2019. “Be Wary of Those Who Ask: A Randomized Experiment on the Size and Determinants of the Enumerator Effect.” The World Bank Economic Review, lhy024, https://doi.org/10.1093/wber/lhy024 Dillon,A., Bardasi, E., Beegle, K., and Serneels, P. 2012.“Explaining Variation in Child Labor Statistics.” Journal of Development Economics, 98, 136-147. https://doi.org/10.1016/j.jdeveco.2011.06.002 Dillon, A., Gourlay, S., McGee, K., and Oseni, G. 2019. “Land measurement bias and its empirical implications: evidence from a validation exercise.” Economic Development and Culture Change 67, 595–624. Dillon, A., Karlan, D., Udry, C., and Zinman, J. 2020. “Good Identification, Meet Good Data. World Development 127, 104796. https://doi:10.1016/j.worlddev.2019.104796 Dillon, A. and Edouard Mensah. 2021. “Respondent Biases in Household Surveys” Global Poverty Lab Working paper. Dillon, A., and Rao, L.N. 2021. “Land Measurement Bias: Comparisons from Global Positioning System, Self-Reports, and Satellite Data.” Global Poverty Lab Working Paper. Dixon-Mueller, R. B., and Anker, R. 1988. “Assessing Women’s Economic Contributions to Development.” World Employment Programme Background Paper 6. Donaldson, D., and Storeygard, A. 2016. “The View from Above: Applications of Satellite Data in Economics.” Journal of Economic Perspectives 30, 171-98. Doss, C., and Kieran, C. 2014. “Standards for Collecting Sex-Disaggregated Data for Gender Analysis; a Guide for CGIAR Researchers.” CGIAR Gender Research Network. Doss, C., Kieran, C., and Kilic, T. 2019. “Measuring Ownership, Control, and use of Assets. ” Feminist Economics 0, 1–25. https://doi.org/10.1080/13545701.2019.1681591 Fabregas, R., Kremer, M., and Schilbach, F. 2019. “Realizing the Potential of Digital Development: The Case of Agricultural Advice.” Science 366. https://doi.org/10.1126/science.aay3038 Fafchamps, M., McKenzie, D., Quinn, S., and Woodruff, C. 2012. “Using PDA Consistency Checks to Increase the Precision of Profits and Sales Measurement in Panels.” Journal of Development Economics 98, 51–57. FAO. 1982. “Estimation of Crop Areas and Yields in Agricultural Statistics.” FAO Economic and Social Development Paper, 22. Rome: FAO. FAO. 2011. World Livestock 2011 – Livestock in food security. Rome: FAO. FAO. 2013. Children’s work in the livestock sector: Herding and beyond. Rome: FAO. FAO. 2015a. World Census of Agriculture 2020.Volume 1: Programme, Concepts and Definitions. Rome: FAO. FAO. 2015b. Handbook on Master Sampling Frames for Agricultural Statistics. Rome: FAO FAO; The World Bank; UN-Habitat. 2019. Measuring Individuals’ Rights to Land: An Integrated Approach to Data Collection for SDG Indicators 1.4.2 and 5.a.1. Washington, DC: World Bank. © FAO, The World Bank, and UN-Habitat. License: CC BY-NC-SA 3.0 IGO References 47 Fermont, A., and Benson, T. 2011. “Estimating yield of food crops grown by smallholder farmers”. IFPRI Discussion Paper 01097. International Food Policy Research Institute, Washington, DC. Fisher, R. 1926. “The Arrangement of Field Experiments.” Journal of the Ministry of Agriculture 33, 503-515. Fitzgerald, J., Gottschalk, P., and Moffitt, R. 1998. “An Analysis of Sample Attrition in Panel Data: The Michigan Panel Study of Income Dynamics.” The Journal of Human Resources 33, 251–299. Fowler Jr, F. J. 1995. Improving survey questions: Design and evaluation. Sage. Fox, K. A. 1986. “Agricultural Economists as World Leaders in Applied Econometrics, 1917–33.” American Journal of Agricultural Economics 68, 381-386. https://doi.org/10.2307/1241449 Friedman, J., Gaddis, I., Kilic, T., Martuscelli, A., Palacios-Lopez, A., and Zezza, A. Forthcoming. “The Distribution of Effort: Physical Activity, Gender Roles, and Bargaining Power in Malawi.” Gaddis, I., Siwatu, G. O., Palacios-Lopez, A., and Pieters, J. 2019. “Measuring farm labor: survey experimental evidence from Ghana.” World Bank Policy Research Working Paper 8717. World Bank, Washington, DC. Garlick, R., Orkin, K., Quinn, S., 2020. “Call Me Maybe: Experimental Evidence on Frequency and Medium Effects in Microenterprise Surveys. World Bank Economic Review 34, 418–443. https://doi.org/10.1093/wber/lhz021 Gideon, L. 2012. Handbook of Survey Methodology for the Social Sciences. Glennerster, R.,Walsh, C., and Diaz-Martin, L. 2018. “A Practical Guide to Measuring Women’s and Girls Empowerment in Impact Evaluation.” https://www.povertyactionlab.org/practical-guide-measuring-womens-and-girls-empowerment- impact-evaluations Global Strategy to improve Agricultural and Rural Statistics. 2017. Handbook on the Agricultural Integrated Survey. AGRIS. Goldstein, M. and Udry, C. 1999. Agricultural innovation and risk management in Ghana. Unpublished Final Report IFPRI. Gourlay, S., Aynekulu, E., Carletto, C., and Shepherd, K. 2017. Spectral Soil Analysis and Household Surveys. Washington, DC: World Bank. Gourlay, S., Kilic, T., and Lobell, D. B. 2019. “A New Spin on an Old Debate: Errors in Farmer-Reported Production and their Implications for Inverse Scale-Productivity Relationship in Uganda.” Journal of Development Economics 141, 102376. Grace, D., Mutua, F., Ochungo, P., Kruska, R., Jones, K., Brierley, L., and Ogutu, F. 2012. “Mapping of poverty and likely zoonoses hotspots.” Zoonoses Project 4, 1-119. Grace, K., Brown, M., and McNally, A. 2014. “Examining the Link Between Food Prices and Food Insecurity: A Multi-level Analysis of Maize Price and Birthweight in Kenya.” Food Policy 46, 56-65. Grosh, M. and Glewwe, P. 2000. Designing household survey questionnaires for developing countries. Washington DC: World Bank. Guarcello, L., Lyon, S., and Rosati, F. C. 2008. “Child Labor and Education for All: An Issue Paper.” The journal of the History of Childhood and Youth 1, 254–266. Heffetz, O., and Ligett, K. 2014. “Privacy and Data-Based Research.” Journal of Economic Perspectives 28(2), 75-98. Hengl, T., Heuvelink, G.B., Kempen, B., Leenaars, J.G., Walsh, M.G., Shepherd, K.D., Sila, A., MacMillan, R.A., Mendes de Jesus, J., Tamene, L. and Tondoh, J.E. 2015. “Mapping Soil Properties of Africa at 250 m Resolution: Random Forests Significantly Improve Current Predictions.” PloS one, 10(6), e0125814. Herberich, D. H., Levitt, S. D., and List, J. A. 2009. “Can Field Experiments Return Agricultural Economics to the Glory Days?” American Journal of Agricultural Economics 91, 1259-1265. https://doi.org/10.1111/j.1467-8276.2009.01294.x Hess, J., Moore, J., Pascale, J., Rothgeb, J., and Keeley, C. 2001. “The Effects of Person-Level versus Household-Level Questionnaire Design on Survey Estimates and Data Quality.” Public Opinion Quarterly 65, 574-584. http:// doi:10.1086/323580 Hill, Z. 2004. “Reducing attrition in panel studies in developing countries.” International Journal of Epidemiology 33, 493- 498. http://doi:10.1093/ije/dyh060 48 AGRICULTURAL SURVEY DESIGN Himelein, K. 2016. “Interviewer Effects in Subjective Survey Questions: Evidence from Timor-Leste.” International Journal of Public Opinion Research 28(4): 511-533 Hoddinott, J., Headey, D., and Dereje, M. 2015. “Cows, Missing Milk Markets, and Nutrition in Rural Ethiopia.” The Journal of Development Studies 51, 958-975. Holden, S., Ali, D., Deininger, K., and Hilhorst, T. 2016. A Land Tenure Module for LSMS. Centre for Land Tenure Studies Working Paper 1/16. Norwegian University of Life Sciences. Husak, G., and Grace, K. 2016. “In Search of a Global Model of Cultivation: Using Remote Sensing to Examine the Characteristics and Constraints of Agricultural Production in the Developing World.” Food Security 8. Iarossi, G. 2006. The Power of Survey Design : A User’s Guide for Managing Surveys, Interpreting Results, and Influencing Respondents. Washington, DC:The World Bank. Jacobs, K., Kes, A. 2015. “The Ambiguity of Joint Asset Ownership: Cautionary Tales from Uganda and South Africa.” The Feminist Economics 21, 23–55. Jain, Meha. 2020. “The Benefits and Pitfalls of Using Satellite Data for Causal Inference”. Review of Environmental Economics and Policy 14(1): 157-169. Janzen, S.A. 2018. “Child Labour Measurement: Whom Should We Ask?” International Labour Review 157, 169–191. Just, R. E., and Pope, R. D. 2001.“The agricultural producer:Theory and statistical measurement.” Chapter 12 of Agricultural Production (Vol. 1, p. 629-741). Elsevier. https://doi.org/10.1016/S1574-0072(01)10015-0 Keita, N. and Carfagna, E. 2009. “Use of Modern Geo-Positioning Devices in Agricultural Censuses and Surveys: Use of GPS for Crop Area Measurement.” Bulletin of the International Statistical Institute, the 57th Session, 2009, Proceedings, Special Topics Contributed Paper Meetings (STCPM22), Durban. Kilic, T., Koolwal, G.B., and Moylan, H.G. 2020. “Are You Being Asked? Impacts of Respondent Selection on Measuring Employment.” World Bank Policy Research Working Paper 9152. World Bank,Washington, DC. Kilic, T., Moylan, H., and Koolwal, G. 2020b. “Getting the (Gender-Disaggregated) Lay of the Land: Impact of Survey Respondent Selection on Measuring Land Ownership and Rights.” World Bank Policy Research Working Paper 9151. World Bank,Washington, DC. Kilic, T., and Moylan, H. 2016. “Methodological Experiment on Measuring Asset Ownership from a Gender Perspective (MEXA).” World Bank,Washington, DC. Kilic, T., Moylan, H.G., Ilukor, J., Mtengula, C., and Pangapanga-Phiri, I. 2018. “Root for the Tubers: Extended-Harvest Crop Production and Productivity Measurement in Surveys.” World Bank Policy Research Working Paper 8618. World Bank,Washington, DC. Kilic, T., Serajuddin, U., Uematsu, H., and Yoshida, N. 2017. “Costing household surveys for monitoring progress toward ending extreme poverty and boosting shared prosperity.” World Bank Policy Research Working Paper 7951. World Bank,Washington, DC. Kilic, T., Van den Broeck, G., Koolwal, G., and Moylan, H. 2020a. “Are You Being Asked? Impacts of Respondent Selection on Measuring Employment.” World Bank Policy Research Working Paper 9152. World Bank,Washington, DC. Kilic, T., Zezza, A., Carletto, C., and Savastano, S. 2017). “Missing (ness) in Action: Selectivity Bias in GPS-based Land Area Measurements.” World Development 92, 143-157. Kosmowski, F., Aragaw, A., Kilian, A., Ambel, A., Ilukor, J., Yigezu, B. and Stevenson, J. 2018. “Varietal Identification in Household Surveys: Results from Three Household–based Methods against the Benchmark of DNA Fingerprinting in Southern Ethiopia”, Experimental Agriculture 55, pp. 371–385. Krosnick, J. A. 1999. “Survey research.” Annual review of psychology, 50, 537–567. Laajaj, R., and Macours, K. 2019. “Measuring Skills in Developing Countries.” Journal of Human Resources, 1018-9805R1. Lobell, D. B., Cassman, K. G., and Field, C. B. 2009. “Crop Yield Gaps: Their Importance, Magnitudes, and Causes.” Annual Review of Environment and Resources, 34, 179-204. http://doi:10.1146/annurev.environ.041008.093740 Lobell, D. B., Di Tommaso, S.,You, C.,Yacoubou Djima, I., Burke, M., and Kilic, T. 2020. “Sight for Sorghums: Comparisons of Satellite-and Ground-Based Sorghum Yield Estimates in Mali.” Remote Sensing, 12(1), 100. References 49 Lobell, D., Azzari, G., Burke, M., Gourlay, S., Jin, Z., Kilic, T., and Murray, S. 2019. “Eyes in the Sky, Boots on the Ground: Assessing Satellite- and Ground-Based Approaches to Crop Yield Measurement and Analysis.” American Journal of Agricultural Economics aaz051, https://doi.org/10.1093/ajae/aaz051 Manski, C. F., and Molinari, F. 2008. “Skip Sequencing: A Decision Problem in Questionnaire Design.” The Annals of Applied Statistics 2, 264–285. Michelson, H., Fairbairn, A., Ellison, B., Maertens, A., and Manyong, V. 2021. “Misperceived Quality: Fertilizer in Tanzania.” Journal of Development Economics 148, 102579. https://doi.org/10.1016/j.jdeveco.2020.102579 Moore, J. C. 1988. “Self/proxy Response Status and Survey Response Quality.” Journal of Official Statistics 4, 155–172. O’Sullivan, M., Rao, A., Raka, B., Kajal, G., and Margaux,V. 2014. Levelling the Field: Improving Opportunities for Women Farmers in Africa. Washington, DC: World Bank. Oseni, G., Durazo, J., and McGee, K. 2017.The use of non-standard measurement units for the collection of food quantity: a guidebook for improving the measurement of food consumption and agricultural production in living standards surveys. LSMS Guidebook, Washington, DC: World Bank. Palacios-Lopez, A., Christiaensen, L., Kilic, T. 2017. “How much of the labor in African agriculture is provided by women?” Food Policy 67, 52–63. https://doi.org/10.1016/j.foodpol.2016.09.017 Payne, S. L. 2014. The Art of Asking Questions: Studies in Public Opinion, 3 (Vol. 451). Princeton University Press. Pica-Ciamarra, U., Baker, D., Morgan, N., Zezza, A., Azzarri, C., Ly, C., Nsiima, L., Nouala, S., Okello, and P., Sserugga, J. 2014. “Investing in the livestock sector: Why Good Numbers Matter, a Sourcebook for Decision Makers on How to Improve Livestock Data.” World Bank. Rozelle, S. 1991. Rural household data collection in developing countries: Designing instruments and methods for collecting farm production data. Technical report. World Bank: Washington, DC. Sagesaka, A., Palacios-Lopez, A., and Amankwah, A. 2020. Measuring Work on Household-Farms using Household Surveys. A Practical Guidebook on Designing Household Surveys for Effective Data Collection of Work on Household Farms. Technical report.World Bank: Washington, DC. Schøning, P. 2005. Handheld GPS Equipment for Agricultural Statistics Surveys: Experiments on Area-Measurements Done during Fieldwork for the Uganda Pilot Census of Agriculture, 2003. Statistisk sentralbyrå. Schündeln, M. 2018. “Multiple Visits and Data Quality in Household Surveys.” Oxford Bulletin of Economics and Statistics 80, 380–405. Serneels, P., Beegle, K., and Dillon, A. 2017. “Do returns to education depend on how and whom you ask?” Economics of Education Review 60, 5–19. Singh, I., Squire, L., and Strauss, J. 1986. Agricultural household models: Extensions, applications, and policy. Washington, DC:World Bank. Sudman, S., and Bradburn, N.M. 1973. “Effects of Time and Memory Factors on Response in Surveys.” Journal Of American Statistics Association 68, 805–815. Thomas, D., Witoelar, F., Frankenberg, E., Sikoki, B., Strauss, J., Sumantri, C., and Suriastini, W. 2012. “Cutting the Costs of Attrition: Results from the Indonesia Family Life Survey.” Journal of Development Economics, 98, 108-123. https:// doi:10.1016/j.jdeveco.2010.08 Udry, C. 1996. “Gender, Agricultural Production, and the Theory of the Household.” Journal of Political Economy, 104(5), 1010-1046. United Nations Statistics Division. 2019. Guidelines for Producing Statistics on Asset Ownership from a Gender Perspective. https://unstats.un.org/edge/publications/docs/Guidelines_final.pdf Wineman, A., Anderson, C. L., Reynolds, T., and Biscaye, P. 2019. “Methods of Crop Yield Measurement on Multi-Cropped Plots: Examples from Tanzania,” Food Security 11,1257–1273. https://doi.org/10.1007/s12571-019-00980-5 Willis, K. J., Seddon, A. W., Long, P. R., Jeffers, E. S., Caithness, N., Thurston, M., and Macias-Fauria, M. 2015. “Remote Assessment of Locally Important Ecological Features across Landscapes: how Representative of Reality?” Ecological Applications 25, 1290-1302. https://doi.org/10.1890/14-1431.1 50 AGRICULTURAL SURVEY DESIGN Wineman, A., Njagi, T., Anderson, C.L., Reynolds, T.W., Alia, D.Y., Wainaina, P., Njue, E., Biscaye, P. and Ayieko, M.W. 2020. “A Case of Mistaken Identity? Measuring Rates of Improved Seed Adoption in Tanzania using DNA Fingerprinting.” Journal of Agricultural Economics,71(3),719-741. Witoelar, F. 2011. Tracking in Longitudinal Household Surveys. Washington, DC: The World Bank, Development Research Group, LSMS-ISA Integrated Surveys on Agriculture. Wollburg, P., Tiberti, M., and Zezza, A. 2020. “Recall Length and Measurement Error in Agricultural Surveys.” Food Policy 102003. https://doi.org/10.1016/j.foodpol.2020.102003 Working, H. 1925. “The Statistical Determination of Demand Curves.” The Quarterly Journal of Economics 39, 503-543. https://doi:10.2307/1883264 Zezza, A., Federighi, G., Kalilou, A. A., and Hiernaux, P. 2016a. “Milking the Data: Measuring Milk Off-Take in Extensive Livestock Systems. Experimental Evidence from Niger.” Food policy 59, 174-186. Zezza, A, Pica-Ciamarra, U, Mugera, HK, Mwisomba,T, and Okello, P. 2016b. Measuring the Role of Livestock in the Household Economy: A Guidebook for Designing Household Survey Questionnaires (English). LSMS Guidebook Washington, D.C. World Bank. http://documents.worldbank.org/curated/en/108351587037911099/ Measuring-the-Role-of-Livestock-in-the-Household-Economy-A-Guidebook-for-Designing-Household- Survey-Questionnaires Zwane, A.P., Zinman, J., Dusen, E.V., Pariente, W., Null, C., Miguel, E., Kremer, M., Karlan, D.S., Hornbeck, R., Giné, X., Duflo, E., Devoto, F., Crepon, B., and Banerjee, A. 2011. Being surveyed can change later behavior and related parameter estimates. Proceedings of the National Academy of Sciences 108, 1821–1826. https:// doi.org/10.1073/pnas.1000776108 51 Appendix I. Glossary TERM DESCRIPTION Administrative data Data generated as a by-product of registration and other administrative and service delivery functions of government entities or other organizations. Examples include registries of large farms or information collected through agricultural extension services. Agricultural season Agricultural season refers to the portion of a year in which seasonal crops are typically grown in a given locale.This is dictated by local climatic conditions. In agricultural survey instruments, the ‘reference agricultural season’ is used to demarcate and refer to the production cycle of interest. Agro-forestry Land use system in which trees or shrubs are deliberately grown on a plot of land together with crops and animals. Area sampling Sampling method used when no complete sampling frame is available, based on dividing the land area under study into smaller areas and sampling from the list of those smaller areas. Attrition In longitudinal surveys, attrition refers to the loss of survey participants (households, holdings, individuals) over time, with each subsequent round of the survey. Bias In statistics, an estimator is biased when its expected value differs from the true value of the underlying parameter. Census The full enumeration or count of an entire population of interest, such as farms or agricultural holdings (agricultural census) or households (population and housing census). Compass and rope  A method to measure the area of a unit of land reliably using poles, robes, compasses. Computer Assisted A survey mode in which the survey questionnaire is pre-programmed on a tablet or Personal Interviewing computer and administered to the respondent in person by an interviewer. (CAPI) Consumption aggregate Total value of items consumed by the household over a given reference period, common welfare metric used widely to determine whether a household is considered poor. Crop cutting Crop cutting is a more objective measurement method for crop production than farmer recall, considered the gold standard for measuring crop production and yields. Randomly located subplots within a given plots are harvested and weighted before and after drying. Crop rotation Farming method whereby crops on a plot are grown one after the other with the aim of maintaining fertile soil on the plot. Crop Yield Quantity of crop obtained per unit of land area used. Data collection mode Data collection mode or survey mode is the method used to collect the data. Examples include face-to-face, telephone, mail, and web-based surveys. Enumerator Interviewer in surveys. 52 AGRICULTURAL SURVEY DESIGN TERM DESCRIPTION External validity Extent to which results from a study or experiment can be generalized beyond its original context. Geospatial data Data related by geographic location, such as maps, satellite or remote sensing data. Gold standard In survey design, a method for collecting data on a given parameter that is regarded as yielding highly reliable data. Head of the household Individual household member who is (considered by other household members) the main decision maker in a given household. Holding Economic unit of agricultural production under single management, comprising all livestock kept and all land used wholly or partly for agricultural production purposes, without regard to title, legal form, or size (FAO, WCA 2020). Household A household usually refers to a group of individuals who eat meals together and live in the same dwelling. In addition, according to the production-based definition of the household, it consists of adult individuals co-mingle revenues from agricultural production and non-farm enterprise for consumption. The production-based definition is used less frequently. Household member Individual who is a member of a given household. Intercropping Cultivation of more than one temporary and/or permanent crop simultaneously on the same plot. Internal validity Extent to which a causal effect is soundly identified in a study or experiment. Land tenure Land tenure defines the property and use rights of an individual or group with respect to land. Longitudinal survey A panel or longitudinal survey is a survey in which the same survey units (households, holdings, individuals) are re-visited for at least one round after the baseline survey round. Measurement error Deviation of a measured value of a parameter from its true value. Multi-topic household Sample survey with the household as its main unit of analysis and that studies multiple survey issues affecting the household, such as education, employment, poverty, migration, among others. National Statistical Office Government agency in charge of producing official statistics. (NSO) Nonresponse In surveys, nonresponse refers to selected respondents altogether failing to participate in the survey (unit nonresponse) or failing to respond to certain questions of the survey (item nonresponse). Non-sampling error Survey error not due to sampling, such as response errors, when survey respondents misreport on survey questions for example because survey questions are poorly framed, or errors resulting from nonresponse. Panel survey A panel or longitudinal survey is a survey in which the same survey units (households, holdings, individuals) are re-visited for at least one round after the baseline survey round. Paper Assisted Personal Survey interview mode whereby the enumerator uses pen and paper to fill the interviewing (PAPI) questionnaire. Parcel A parcel is a piece of land of one land tenure type entirely surrounded by other land, water, road, forest or other features not forming part of the holding, or forming part of the holding under a different land tenure type.A parcel may comprise one or more plots. Appendix I. Glossary 53 TERM DESCRIPTION Pastoralism Livestock production system involving the grazing of livestock herds on rangelands and large pastures typically practiced by nomadic people who move with their herds. Permanent crops Crops with a growing cycle of more than one year (FAO, WCA 2020), and which do not need to be replanted after one growing cycle. Also referred to as perennial crops. Plot A plot is defined as a continuous piece of land on which a specific crop or a mixture of crops is grown, or which is fallow or waiting to be planted. Plot manager The manager of an agricultural plot is typically the main decision maker regarding crops planted, inputs applied, and output harvested. Post-harvest questionnaire Questionnaire administered during the post-harvest visit of the survey, that is, a survey visit taking place after the farm’s main harvest of temporary/seasonal crops has been completed. Post-planting questionnaire Questionnaire administered during the post-planting visit of the survey, that is a survey visit taking place after the farm has completed the planting period of its main temporary/ seasonal crops. Productivity In this volume, productivity refers to agricultural productivity, whether yield (output per area of land), labor productivity (output per unit of labor or per worker), or total factor productivity, taking into account all inputs and outputs. Proxy response In proxy response, an individual responds on behalf another individual, rather than all individuals responding directly for themselves. Randomized controlled Scientific experiment to test the causal effect of a treatment or an intervention while trial controlling for confounding factors through experimental design. Recall Recall refers to the process of respondents self-reporting of past events of interest. Recall period Length of the period between the interview and an event the respondent is requested to recall. Reference period The period of reference for a survey question, for example, the reference agricultural season or the last 7 days. Representativeness A sample is representative if it accurately reflects the study population of interest with respect to a set of predefined characteristics. Research design Methods, techniques, and strategies chosen to answer a given research question. Root crops Root crops or root vegetables are those crops whose roots are meant for consumption, such as cassava or onions. Sample survey Survey based on a sample of the entire population of interest, rather than full enumeration, usually designed to be representative of the population of interest. Sampling Selecting a predetermined number of units from a population of interest, such as farms or households. Sampling error Deviation of sample means from true population means leading to reduced representativeness of the sample. Sampling frame A list of units, such as farms, from which a survey sample is drawn. Sex-disaggregated data Data that allows distinguishing between women and men at the level of the individual. Skip sequences/patterns In survey questionnaires, a skip pattern routes the respondent to a specific question based on his or her answer to a previous question. 54 AGRICULTURAL SURVEY DESIGN TERM DESCRIPTION Smallholder A smallholder or smallholder farmer is a person who owns and/or operates a small- scale agricultural holding, whether defined by land area, production volume, production technology, or other factors. Social desirability bias Social desirability bias refers to a response bias whereby survey respondents answer survey questions in a way they perceive as pleasing the interviewer. Survey design Methods and techniques for developing and implementing a survey. Survey experiment Scientific experiment to test the impact of survey design on measurement through surveys. Survey methodology The study of survey methods. Survey respondent Individual who respondents to survey questions. Temporary crops Crops with at least one growing cycle (planting and harvesting) per agricultural year, often once per reference agricultural season. Total survey error The sum of all errors from design, implementation, data processing and analysis of surveys. Tropical livestock units A standardized unit for measuring the stock of live animals independent of their breed or size. Unit of analysis The unit of analysis or unit of observation is the unit of interest of a study or survey, such as farms, individuals, households, or enterprises. The unit of analysis may vary in one survey or study. Welfare Measure of material wellbeing. Appendix II. Agricultural Reference Questionnaire 55 Appendix II. Agricultural Reference Questionnaire Agricultural Survey Design Lessons from the LSMS-ISA and Beyond Editable versions of the Agricultural Survey Design reference questionnaires are available for download at: https://www.worldbank.org/en/programs/lsms/publication/AgriculturalSurveyDesign 56 AGRICULTURAL SURVEY DESIGN LSMS-ISA REFERENCE QUESTIONNAIRE THIS INFORMATION IS STRICTLY CONFIDENTIAL AND IS TO BE USED FOR STATISTICAL PURPOSES ONLY. POST-PLANTING AGRICULTURE QUESTIONNAIRE HOUSEHOLD IDENTIFICATION CODE NAME 1. DISTRICT: ...................................................................................................................................................................................................................... 2. STATE: ...................................................................................................................................................................................................................... 3. ENUMERATION AREA: 4.VILLAGE NAME: ...................................................................................................................................................................................................................... 5. HOUSEHOLD ID (FROM LIST): 6. NAME OF HOUSEHOLD HEAD: ...................................................................................................................................................................................................................... LATITUDE (S) LONGITUDE (E) 7. GPS COORDINATES OF DWELLING __ __ O __ __ . __ __ __ __ __ __ O __ __ . __ __ __ SURVEY STAFF DETAILS 8. ENUMERATOR CODE: ...................................................................................................................................................................................................................... 9. DATE OF INTERVIEW DD MM YYYY 10. SUPERVISOR CODE: ...................................................................................................................................................................................................................... 11. DATE OF INSPECTION: DD MM YYYY RECORD GENERAL NOTES ABOUT THE INTERVIEW AND ANY SPECIAL INFORMATION THAT WILL BE HELPFUL FOR SUPERVISORS AND DATA ANALYSIS. PLEASE MARK AN `X’ IN BOX IF HOUSEHOLD REFUSAL. PROVIDE DETAILS. Appendix II. Agricultural Reference Questionnaire 57 TABLE OF CONTENTS MODULE 1 Parcel roster MODULE 2 Parcel details MODULE 3 Plot roster MODULE 4 Plot details MODULE 5A Labor inputs (household) MODULE 5B Labor inputs (hired & exchange) MODULE 6 Crop roster MODULE 7 Seed acquisition Network roster Post-planting unit appendix 58 AGRICULTURAL SURVEY DESIGN MODULE 1. PARCEL ROSTER 1. During the [REFERENCE AGRICULTURAL SEASON], did you OR any member of your household own any land OR cultivate any land? Please include land planted with trees. YES.....1 NO.....2 END A PARCEL IS DEFINED AS A PIECE OF LAND EXPLOITED BY ONE OR MORE PERSONS AS A SINGLE FARMING UNIT. A PARCEL MAY BE BOUNDED BY NATURAL BOUNDARIES, AND MAY COMPRISE ONE OR MORE PLOTS. THE NATURAL BOUNDARY OF A FIELD MAY BE A ROAD, A WATERWAY OR A FIELD BELONGING TO ANOTHER FARM. A PLOT IS DEFINED AS A CONTINUOUS PIECE OF LAND ON WHICH A UNIQUE CROP OR A MIXTURE OF CROPS IS GROWN, UNDER A UNIFORM, CONSISTENT CROP MANAGEMENT SYSTEM. PLEASE LIST ALL PARCELS YOU OR ANYONE IN YOUR HOUSEHOLD OWNED OR CULTIVATED DURING THE [REFERENCE AGRICULTURAL SEASON]. PLEASE ALSO INCLUDE PARCELS CULTIVATED ONLY WITH TREE CROPS OR CASSAVA DURING THE [REFERENCE AGRICULTURAL SEASON]. 2 3 4 5 6 7 PARCEL NAME LOCATION & How many What is the distance from [PARCEL] to: What is the area of this [PARCEL]? RECORD THE COORDINATES FOR THE CORNER DESCRIPTION crop plots ASK THE FARMER TO ESTIMATE THE OF THE PLOT AT WHICH YOU STARTED AREA are in this AREA FIRST. MEASURE THE AREA WITH MEASUREMENT. [PARCEL]? THE GPS BEFORE PROCEEDING WITH IF YOU DID NOT RECORD THE GPS COORDINATES, THE REMAINING MODULES. PLEASE SPECIFY REASON. CODES FOR UNIT: CODES FOR REASON: ACRE...................................................... 1 LONG DISTANCE WITHIN THIS DISTRICT.............. 1 HECTARE.............................................. 2 LONG DISTANCE OUTSIDE THIS DISTRICT............ 2 SQUARE METERS............................... 3 HOUSEHOLD REFUSED................................................... 3 YARDS.................................................... 4 OTHER (SPECIFY)............................................................... 4 OTHER (SPECIFY)............................... 5 PARCEL ID a. b. a. b. HOME ROAD MARKET FARMER ESTIMATION GPS MEASURE LATITUDE (S) LONGITUDE (E) REASON km km km AREA UNIT AREA IN ACRES 1 _ _ _ _ ._ _ _ _ _ _ ._ _ _ _ O_ _ ._ _ _ _ _ _ O_ _ ._ _ _ 2 _ _ _ _ ._ _ _ _ _ _ ._ _ _ _ O_ _ ._ _ _ _ _ _ O_ _ ._ _ _ 3 _ _ _ _ ._ _ _ _ _ _ ._ _ _ _ O_ _ ._ _ _ _ _ _ O_ _ ._ _ _ 4 _ _ _ _ ._ _ _ _ _ _ ._ _ _ _ O_ _ ._ _ _ _ _ _ O_ _ ._ _ _ 5 _ _ _ _ ._ _ _ _ _ _ ._ _ _ _ O_ _ ._ _ _ _ _ _ O_ _ ._ _ _ 6 _ _ _ _ ._ _ _ _ _ _ ._ _ _ _ O_ _ ._ _ _ _ _ _ O_ _ ._ _ _ 7 _ _ _ _ ._ _ _ _ _ _ ._ _ _ _ O_ _ ._ _ _ _ _ _ O_ _ ._ _ _ 8 _ _ _ _ ._ _ _ _ _ _ ._ _ _ _ O_ _ ._ _ _ _ _ _ O_ _ ._ _ _ Appendix II. Agricultural Reference Questionnaire 59 MODULE 2. PARCEL DETAILS 1 2 3 4 5 6 ENUMERATOR: How was this [PARCEL] acquired? How much do you pay How many months did How much in total did Under which tenure system is this RECORD THE ID OF the owner for the use this payment cover? you pay for [PARCEL] [PARCEL]? THE RESPONDENT. of this [PARCEL]? (Include both cash and GRANTED BY CUSTOMARY/COMMUNITY LIST FROM AUTHORITIES.......................................................... 1 6 ESTIMATE THE payments in-kind)? CUSTOMARY....................................... 1 HOUSEHOLD ALLOCATED BY GOVERNMENT....................... 2 6 VALUE OF IN- FREEHOLD........................................... 2 ROSTER” ALLOCATED BY FAMILY MEMBER..................... 3 6 KIND PAYMENTS, LEASEHOLD......................................... 3 INHERITED BY THE DEATH EXCLUDING CROP STATE..................................................... 4 OF A FAMILY MEMBER........................................... 4 6 OUTPUTS COMMUNITY/GROUP RIGHT....... 5 PURCHASED............................................................. 5 5 COOPERATIVES.................................. 6 RENTED IN, SHORT-TERM (< 3 YEARS)........... 6 OTHER(SPECIFY)................................ 7 RENTED IN, LONG-TERM.................................... 7 SHARECROPPED IN............................................... 8 6 BORROWED FOR FREE........................................ 9 6 BRIDE PRICE...........................................................10 6 GIFT FROM NON-HOUSEHOLD MEMBER...11 6 MOVED IN WITHOUT PERMISSION...............12 18 PARCEL ID OTHER (SPECIFY)..................................................13 6 CASH IN-KIND PID CODE $ $ NUMBER $ CODE 1 2 3 4 5 6 7 8 60 AGRICULTURAL SURVEY DESIGN MODULE 2. PARCEL DETAILS 7 8 9 10 Who in the household [owns/ In which Does your What type of documents does your household have for this [PARCEL], and which household members are listed as owners or holds use rights to] this year household have a use rights holders on each? [PARCEL]? was this document for this LIST UP TO 3, SHOW PHOTO AID LIST UP TO 4 JOINT OWNERS [PARCEL] [PARCEL] issued OR USE RIGHT HOLDERS acquired? by or registered at FROM HOUSEHOLD ROSTER.” the Land Registry/ DOCUMENT TYPE: Cadastral Agency, TITLE DEED........................................................................................................................1 such as a title CERTIFICATE OF CUSTOMARY OWNERSHIP........................................................2 deed, certificate CERTIFICATE OF OCCUPANCY.................................................................................3 of ownership, CERTIFICATE OF HEREDITARY ACQUISITION LISTED IN REGISTRY............4 certificate of SURVEY PLAN....................................................................................................................5 hereditary RENTAL CONTRACT, REGISTERED...........................................................................6 acquisition, lease or LEASE, REGISTERED.........................................................................................................7 rental contract? OTHER (SPECIFY).............................................................................................................8 YES.....1 IF NO HOUSEHOLD MEMBER ON DOCUMENT, ENTER “98” NO.....2 11 IF DON’T KNOW, ENTER “99” PARCEL ID DOCUMENT #1 DOCUMENT #2 DOCUMENT #3 HHID HHID HHID HHID DOC. HHID HHID HHID HHID DOC. HHID HHID HHID HHID DOC. HHID HHID HHID HHID CODE CODE CODE CODE TYPE CODE CODE CODE CODE TYPE CODE CODE CODE CODE TYPE CODE CODE CODE CODE #1 #2 #3 #4 YEAR #1 #2 #3 #4 #1 #2 #3 #4 #1 #2 #3 #4 1 2 3 4 5 6 7 8 Appendix II. Agricultural Reference Questionnaire 61 MODULE 2. PARCEL DETAILS 11 12 13 14 15 16 ENUMERATOR: Does anyone in the Who can decide whether to sell Does anyone in the Who can decide whether to bequeath Does anyone in the IS THIS PARCEL household have the right to [PARCEL]? household have the right this [PARCEL]? household have the right RENTED IN OR sell [PARCEL], either alone LIST UP TO 4 ID CODES FROM to bequeath this [PARCEL], to use this [PARCEL] as SHARE CROPPED or with someone else? HOUSEHOLD ROSTER AND 1 either alone or with collateral, either alone or IN ACCORDING TO someone else? LIST UP TO 4 ID CODES FROM with someone else? CODE FROM NETWORK ROSTER, IF HOUSEHOLD ROSTER AND 1 QUESTION 2? APPLICABLE. YES...............................1 CODE FROM NETWORK ROSTER, IF NO...............................2 14 YES...............................1 APPLICABLE. YES...............................1 YES.....1 18 DONT’ KNOW..... 98 14 NO...............................2 16 NO...............................2 18 PARCEL ID NO.....2 REFUSAL.................. 99 14 DONT’ KNOW..... 98 16 DONT’ KNOW..... 98 18 REFUSAL.................. 99 16 REFUSAL.................. 99 18 HHID HHID HHID HHID HHID HHID HHID HHID CODE CODE CODE CODE NWID CODE CODE CODE CODE NWID #1 #2 #3 #4 #1 #2 #3 #4 1 2 3 4 5 6 7 8 62 AGRICULTURAL SURVEY DESIGN MODULE 2. PARCEL DETAILS 17 18 19 20 Who in the household has the right to On a scale from 1 to 5, where 1 is not at all likely and 5 is extremely likely, how likely is If you were to sell this If this [PARCEL] were to be sold/rented use the [PARCEL] as collateral, either [NAME of owner/use right holder] to involuntarily lose ownership or use rights to this [PARCEL] today, how out today, who would decide how the alone or jointly with someone else? [PARCEL] in the next 5 years? much could you sell money is used? LIST UP TO 4 ID CODES FROM REFER TO ID CODES IN Q7. IF MOVED IN WITHOUT PERMISSION, ASK ABOUT it for? LIST UP TO 4 MEMBERS FROM HOUSEHOLD ROSTER AND 1 PRINCIPLE COUPLE OF HOUSEHOLD. HOUSEHOLD ROSTER. LIST UP TO 1 CODE FROM NETWORK ROSTER, IF FROM NETWORK ROSTER APPLICABLE. NOT AT ALL LIKELY........................... 1 SLIGHTLY LIKELY................................ 2 MODERATELY LIKELY....................... 3 VERY LIKELY......................................... 4 PARCEL ID EXTREMELY LIKELY........................... 5 HHID HHID HHID HHID INDIVIDUAL 1 INDIVIDUAL 2 INDIVIDUAL 3 INDIVIDUAL 4 HHID HHID HHID HHID CODE CODE CODE CODE NWID $ CODE CODE CODE CODE NWID #1 #2 #3 #4 ID RESPONSE ID RESPONSE ID RESPONSE ID RESPONSE #1 #2 #3 #4 1 2 3 4 5 6 7 8 Appendix II. Agricultural Reference Questionnaire 63 MODULE 2. PARCEL DETAILS 21 22 23 24 25 What are the three main uses of this [PARCEL]? ENUMERATOR: ENUMERATOR: How much do you receive from How many months did this payment is this parcel entirely or is this parcel entirely or renting out/sharecropping out this cover? partially used for crop partially rented out or [PARCEL]? RESIDENTIAL............................................................ 1 production? sharecropped out? ESTIMATE THE VALUE OF IN-KIND CROP PRODUCTION............................................ 2 GRAZING (MEADOWS AND PASTURES......... 3 RECEIPTS, INCLUDING CROP TEMPORARILY FALLOW....................................... 4 OUTPUTS FARM BUILDINGS................................................... 5 AQUACULTURE....................................................... 6 YES.....1 YES.....1 FOREST....................................................................... 7 NO.....2 NO.....2 NEXT PARCEL BUSINESS/COMMERCIAL..................................... 8 UNUSED..................................................................... 9 RENTED OUT/SHARECROPPED OUT...........10 GAVE OUT FOR FREE..........................................11 NO SECOND USE.................................................12 DON’T KNOW......................................................88 PARCEL ID OTHER(SPECIFY....................................................99 CASH IN-KIND 1st 2nd 3rd $ $ 1 2 3 4 5 6 7 8 64 AGRICULTURAL SURVEY DESIGN MODULE 3. PLOT ROSTER A PARCEL IS DEFINED AS A PIECE OF LAND EXPLOITED BY ONE OR MORE PERSONS AS A SINGLE FARMING UNIT. A PARCEL MAY BE BOUNDED BY NATURAL BOUNDARIES, AND MAY COMPRISE ONE OR MORE PLOTS. THE NATURAL BOUNDARY OF A FIELD MAY BE A ROAD, A WATERWAY OR A FIELD BELONGING TO ANOTHER FARM. A PLOT IS DEFINED AS A CONTINUOUS PIECE OF LAND ON WHICH A UNIQUE CROP OR A MIXTURE OF CROPS IS GROWN, UNDER A UNIFORM, CONSISTENT CROP MANAGEMENT SYSTEM. PLEASE LIST ALL PLOTS WITHIN THE PARCELS ENTIRELY OR PARTIALLY USED FOR CROP CULTIVATION DURING THE [REFERENCE AGRICULTURAL SEASON]. 1 2 3 4 5 PLOT NAME LOCATION & During the [REFERENCE AGRICULTURAL What is the area of this [PLOT]? ENUMERATOR: IS THIS DESCRIPTION SEASON], is this [PLOT]… ASK THE FARMER TO ESTIMATE THE AREA. PLOT CULTIVATED OR READ ANSWERS LEFT FALLOW? REFER TO QUESTION 3. CODES FOR UNIT: CULTIVATED.............................................. 1 ACRE............................ 1 RENTED OUT........................................... 2 YES.....1 GIVEN OUT FOR FREE........................... 3 HECTARE.................... 2 SQUARE METERS..... 3 NO.....2 NEXT PLOT FALLOW...................................................... 4 FOREST /WOODLOT............................. 5 YARDS.......................... 4 PASTURE..................................................... 6 OTHER (SPECIFY)..... 5 SHARECROPPED OUT........................... 7 PARCEL ID OTHER (SPECIFY)................................. -99 PLOT ID FARMER ESTIMATION AREA UNIT 1 1 _ _ _ _ ._ _ 1 2 _ _ _ _ ._ _ 1 3 _ _ _ _ ._ _ 1 4 _ _ _ _ ._ _ 1 … _ _ _ _ ._ _ 2 1 _ _ _ _ ._ _ 2 2 _ _ _ _ ._ _ 2 3 _ _ _ _ ._ _ 2 4 _ _ _ _ ._ _ … … _ _ _ _ ._ _ Appendix II. Agricultural Reference Questionnaire 65 MODULE 3. PLOT ROSTER A PARCEL IS DEFINED AS A PIECE OF LAND EXPLOITED BY ONE OR MORE PERSONS AS A SINGLE FARMING UNIT. A PARCEL MAY BE BOUNDED BY NATURAL BOUNDARIES, AND MAY COMPRISE ONE OR MORE PLOTS. THE NATURAL BOUNDARY OF A FIELD MAY BE A ROAD, A WATERWAY OR A FIELD BELONGING TO ANOTHER FARM. A PLOT IS DEFINED AS A CONTINUOUS PIECE OF LAND ON WHICH A UNIQUE CROP OR A MIXTURE OF CROPS IS GROWN, UNDER A UNIFORM, CONSISTENT CROP MANAGEMENT SYSTEM. PLEASE LIST ALL PLOTS WITHIN THE PARCELS ENTIRELY OR PARTIALLY USED FOR CROP CULTIVATION DURING THE [REFERENCE AGRICULTURAL SEASON]. 6 7 8 9 10 MEASURE THE AREA RECORD THE COORDINATES FOR THE CORNER OF THE RECORD NUMBER OF RECORD GPS RECORD THE WEATHER CONDITIONS AT TIME OF THE PLOT USING PLOT AT WHICH YOU STARTED AREA MEASUREMENT. SATELLITES GPS TRACKED ACCURACY OF MEASUREMENT THE GPS DEVICE. IF YOU DID NOT RECORD THE GPS COORDINATES, TO CAPTURE PLOT PLEASE SPECIFY REASON. COORDINATES CLEAR/SUNNY............................................................. 1 MOSTLY CLEAR / MOSTLY SUNNY....................... 2 CODES FOR REASON: PARTLY CLOUDY / PARTLY SUNNY....................... 3 MOSTLY CLOUDY / LONG DISTANCE WITHIN THIS DISTRICT........... 1 CONSIDERABLE CLOUDINESS............................... 4 LONG DISTANCE OUTSIDE THIS DISTRICT......... 2 COMPLETELY CLOUDY............................................. 5 HOUSEHOLD REFUSED................................................ 3 RAINY.............................................................................. 6 PARCEL ID OTHER (SPECIFY............................................................. 4 PLOT ID GPS MEASURE a. b. AREA IN ACRES LATITUDE (S) LONGITUDE (E) REASON O O 1 1 _ _ _ _ ._ _ _ _ _ _ ._ _ _ _ _ _ _ ._ _ _ 1 2 _ _ _ _ ._ _ _ _ O_ _ ._ _ _ _ _ O_ _ ._ _ _ 1 3 _ _ _ _ ._ _ _ _ O_ _ ._ _ _ _ _ O_ _ ._ _ _ 1 4 _ _ _ _ ._ _ _ _ O_ _ ._ _ _ _ _ O_ _ ._ _ _ O 1 … _ _ _ _ ._ _ _ _ _ _ ._ _ _ _ _ O_ _ ._ _ _ O 2 1 _ _ _ _ ._ _ _ _ _ _ ._ _ _ _ _ O_ _ ._ _ _ O 2 2 _ _ _ _ ._ _ _ _ _ _ ._ _ _ _ _ O_ _ ._ _ _ O 2 3 _ _ _ _ ._ _ _ _ _ _ ._ _ _ _ _ O_ _ ._ _ _ O 2 4 _ _ _ _ ._ _ _ _ _ _ ._ _ _ _ _ O_ _ ._ _ _ O … … _ _ _ _ ._ _ _ _ _ _ ._ _ _ _ _ O_ _ ._ _ _ 66 AGRICULTURAL SURVEY DESIGN MODULE 4. PLOT DETAILS LIST ALL PLOTS ARE THAT NOT "WOODLOT" OR "PASTURE" ACCORING TO QUESTION 3 IN MODULE 3. 1 2 3 4 5 6 7 Who in the household Are there other Who are the other decision ENUMERATOR: What is the predominant How would you rate the What are the causes of these makes the decisions household makers on the [PLOT]? RECORD THE soil type of this [PLOT]? extent of erosion on this erosion problems? concerning crops to be members that the LIST UP TO 2 HOUSEHOLD ID OF THE READ ANSWERS [PLOT]? READ ANSWERS planted, input use and primary decision MEMBERS FROM RESPONDENT. READ ANSWERS the timing of cropping maker consults LIST UP TO 2 HOUSEHOLD ROSTER LIST FROM activities on this regarding crop SANDY................................... 1 HOUSEHOLD [PLOT]? choice, input use LOAM BETWEEN NO EROSION...............1 8 TERRAIN............................... 1 ROSTER” SANDY & CLAY...................2 and timing of LOW................................2 FLOODING.......................... 2 THIS DOES cropping activities CLAY....................................... 3 MODERATE...................3 NOT NEED TO WIND..................................... 3 on this [PLOT]? OTHER(SPECIFY)............ -99 HIGH...............................4 CORRESPOND TO ANIMALS............................... 4 PARCEL ID THE OWNER OTHER (SPECIFY)........... -99 PLOT ID YES.....1 NO.....2 4 PID PID PID PID 1ST 2ND 1 1 1 2 1 3 1 4 1 … 2 1 2 2 2 3 Appendix II. Agricultural Reference Questionnaire 67 MODULE 4. PLOT DETAILS LIST ALL PLOTS ARE THAT NOT "WOODLOT" OR "PASTURE" ACCORING TO QUESTION 3 IN MODULE 3. 8 9 10 11 12 13 14 What type of erosion control/water What is the slope of Is this [PLOT] Is there any What is the method of irrigating What is the source of ENUMERATOR: harvesting facilities are on this [PLOT]? this [PLOT]? swamp/ wetland? system of plants/pouring water on the plants on water on this [PLOT]? REFER TO irrigation on this this [PLOT]? QUESTION 5. WAS READ ANSWERS READ ANSWERS READ ANSWERS THIS PLOT LEFT [PARCEL]? READ ANSWERS LIST UP TO TWO STRUCTURES. IF YES.....1 FALLOW DURING NONE, RECORD “1” IN BOTH FLAT............................ 1 NO.....2 WELL.......................... 1 THE [REFERENCE YES.....1 AGRICULTURAL SLIGHT SLOPE........... 2 MANUAL IRRIGATION.......................... 1 BOREHOLE................ 2 SEASON]? MODERATE SLOPE.... 3 NO.....2 14 SPRINKLER IRRIGATION....................... 2 LAKE/POND............... 3 NO EROSION CONTROL.......................... 1 TERRACES.................................................... 2 STEEP, HILLY............... 4 DRIP IRRIGATION.................................. 3 RIVER/STREAM........... 4 YES.....1 17 EROSION CONTROL BUNDS.................... 3 FLOODING/SURFACE IRRIGATION..... 4 OTHER(SPECIFY)....... 9 NO.....2 GABIONS / SANDBAGS.............................. 4 EQUIPPED WETLAND/ INLAND VALLEY BOTTOMS.................. 5 VETIVER GRASS........................................... 5 EQUIPPED FLOOD................................. 6 TREE BELTS................................................... 6 RECESSION CULTIVATION................... 7 WATER HARVEST BUNDS.......................... 7 PARCEL ID SPATE IRRIGATION................................ 8 DRAINAGE DITCHES.................................. 8 PLOT ID OTHER(SPECIFY).................................... 9 OTHER (SPECIFY)........................................ 9 1ST 2ND 1 1 1 2 1 3 1 4 1 … 2 1 2 2 2 3 68 AGRICULTURAL SURVEY DESIGN MODULE 4. PLOT DETAILS LIST ALL PLOTS ARE THAT NOT "WOODLOT" OR "PASTURE" ACCORING TO QUESTION 3 IN MODULE 3. 15 16 17 18 19 20 What was the most recent For how Why was this [PLOT] left fallow? ENUMERATOR: How did you prepare land for planting What implements/equipment did year in which this [PLOT] many years IS THIS [PLOT] on [PLOT] during the [REFERENCE you use to prepare land for planting was it left LIST UP TO 2 “CULTIVATED” AGRICULTURAL SEASON]? on [PLOT] during the [REFERENCE was left fallow? fallow? ACCORDING TO AGRICULTURAL SEASON]? IF NEVER RECORD GOOD FOR LAND..................................... 1 QUESTION 5? NO-TILLAGE................................................ 1 LIST UP TO 2. ZERO. LACK OF NON-LABOUR INPUTS............. 2 MULCH TILLAGE......................................... 2 IF DON’T KNOW, LACK OF HOUSEHOLD LABOUR............. 3 HAND HOE.................................................. 1 RECORD -9999 STRIP OR ZONAL TILLAGE........................ 3 LACK OF HIRED LABOUR......................... 4 ANIMAL, MOULDBOARD PLOUGH.......... 2 RIDGE TILL (INCLUDING NO-TILL OR IF NEVER OR DON’T LACK OF EQUIPMENT............................... 5 ANIMAL, DISC PLOUGH............................. 3 RIDGES......................................................... 4 KNOW (-9999),  35. LACK OF CREDIT........................................ 6 ANIMAL, RIPPER........................................... 4 REDUCED OR MINIMUM TILLAGE............ 5 OTHER (SPECIFY)........................................ 7 TRACTOR, MOULDBOARD PLOUGH...... 5 MECHANISED SYSTEM............................... 6 HAVE ENOUGH LAND............................... 8 TRACTOR, DISC PLOUGH......................... 6 PARCEL ID TRADITIONAL TILLAGE............................. 7 OTHER (SPECIFY)........................................ 8 TRACTOR, RIPPER....................................... 7 PLOT ID OTHER (SPECIFY)........................................ 8 4-DIGIT-YEAR 1ST 2ND IMPLEMENT #1 IMPLEMENT #2 1 1 1 2 1 3 1 4 1 … 2 1 2 2 2 3 Appendix II. Agricultural Reference Questionnaire 69 MODULE 5A. LABOR INPUTS (HOUSEHOLD) LIST ALL CULTIVATED PLOTS 1 2 3 4 ENUMERATOR: Has [NAME] worked on [PLOT] for On how many days How many hours per day What activities has [NAME] performed on [PLOT] during the RECORD THE ID OF any activity during the [REFENCE has [NAME] worked on average did [NAME] [REFENCE AGRICULTURAL SEASON]? THE RESPONDENT. AGRICULTURAL SEASON]? on [PLOT] during work on [PLOT]? SELECT ALL THAT APPLY the [REFENCE AGRICULTURAL YES.....1 LAND PREPARATION.................................................................................. 1 SEASON]? NO.....2 NEXT PERSON - PLOT INDIVIDUAL ID PLANTING.................................................................................................... 2 WEEDING..................................................................................................... 3 RIDGING, FERTILIZING, OTHER NON-HARVEST ACTIVITIES................. 4 PARCEL ID HARVESTING................................................................................................ 5 PLOT ID SUPERVISION................................................................................................ 6 PID DAYS (TOTAL) HOURS PER DAY CODES 1 1 1 1 1 2 1 1 3 1 1 4 1 1 … 1 2 1 1 2 2 1 2 3 1 2 4 … … … 70 AGRICULTURAL SURVEY DESIGN MODULE 5B. LABOR INPUTS (HIRED & EXCHANGE) HIRED LABOR 1 2 3 4 5 6 Since your Since your Since your How many hours Normally, how much What activities did [PERSON TYPE] perfom on [PLOT]? household started household started household started per day did a typical did your household SELECT ALL THAT APPLY preparing [PLOT] preparing [PLOT] preparing [PLOT] hired [PERSON pay per day to the for [REFERENCE for [REFERENCE for [REFERENCE TYPE] work on hired [PERSON AGRICULTURAL AGRICULTURAL AGRICULTURAL [PLOT]? TYPE] to work onLAND PREPARATION.......................................................................... 1 SEASON], did you SEASON], how SEASON], how [PLOT]? PLANTING............................................................................................ 2 hire any [PERSON many [PERSON many days did INCLUDE IN-KIND WEEDING............................................................................................. 3 PERSON TYPE TYPE] to work on TYPE] did you or a typical hired RIDGING, FERTILIZING, OTHER NON-HARVEST ACTIVITIES......... 4 [PLOT]? anyone else in the [PERSON TYPE] PAYMENTS HARVESTING........................................................................................ 5 PARCEL ID household hire to work on [PLOT]? RECORD PAYMENT SUPERVISION........................................................................................ 6 PLOT ID work on [PLOT]? PER PERSON PER YES.....1 NO.....2 7 DAY NUMBER DAYS HOURS PER DAY $ PER DAY CODES 1 MEN 1 1 WOMEN 1 1 CHILDREN 1 (UNDER 15) 1 2 MEN 1 2 WOMEN 1 2 CHILDREN (UNDER 15) 1 3 MEN 1 3 WOMEN 1 3 CHILDREN (UNDER 15) … … … Appendix II. Agricultural Reference Questionnaire 71 MODULE 5B. LABOR INPUTS (HIRED & EXCHANGE) EXCHANGE LABOR 7 8 9 10 11 Since your household started preparing this Since your Since your How many hours What activities did [PERSON TYPE] perfom on [PLOT] without [PLOT] for [REFERENCE AGRICULTURAL household started household started per day did a typical pay? SEASON], have any [PERSON TYPE] from preparing [PLOT] preparing [PLOT] [PERSON TYPE] SELECT ALL THAT APPLY other households worked on this [PLOT] for [REFERENCE for [REFERENCE work on [PLOT]? free of charge, as exchange labourers or to AGRICULTURAL AGRICULTURAL assist for nothing in return? SEASON], how SEASON], on how LAND PREPARATION.......................................................................... 1 many [PERSON many days did a PLANTING............................................................................................ 2 PERSON TYPE TYPE] worked on typical [PERSON WEEDING............................................................................................. 3 YES.....1 [PLOT] without pay? TYPE] work wihtout RIDGING, FERTILIZING, OTHER NON-HARVEST ACTIVITIES......... 4 PARCEL ID NO.....2 NEXT PERSON TYPE-PLOT pay on [PLOT]? HARVESTING........................................................................................ 5 PLOT ID SUPERVISION........................................................................................ 6 NUMBER DAYS HOURS PER DAY CODES 1 MEN 1 1 WOMEN 1 1 CHILDREN 1 (UNDER 15) 1 2 MEN 1 2 WOMEN 1 2 CHILDREN (UNDER 15) 1 3 MEN 1 3 WOMEN 1 3 CHILDREN (UNDER 15) … … … 72 AGRICULTURAL SURVEY DESIGN MODULE 6. CROP ROSTER 1 2 3 4 5 Please list all the field crops that you have RECORD THE ID OF During [REFERENCE AGRICULTURAL Is the [CROP] on [PLOT] a ENUMERATOR: SELECT FIELD OR cultivated since the beginning of [REFERENCE THE RESPONDENT. SEASON], approximately what percent purestand or inter-cropped with TREE CROP ON THIS PLOT AGRICULTURAL SEASON] and the tree and of [PLOT] is planted with [CROP]? another crop? permanent crops that you have by [PLOT]. LIST FROM HOUSEHOLD ROSTER ENTIRE PLOT (100).......................................100 PURESTAND.................. 1 FIELD CROP............................ 1 ENUMERATOR: START WITH THE FIRST PLOT THE RESPONDENT (PLOT 3/4 THREE QUARTERS (75%)....................... 75 INTERCROPPED........... 2 TREE/PERMANENT CROP..... 2 12 MANAGER) MANAGES AND RECORD 2/3 TWO THIRDS (66%)................................. 66 EACH CROP, USING A SEPARATE ROW 1/2 HALF (50%)................................................. 50 FOR EACH ONE. THEN ASK ABOUT THE 1/3 ONE THIRD (33%).................................... 33 CROPS ON THE SECOND PLOT THIS 1/4 ONE QUARTER (25%)............................. 25 RESPONDENT MANAGES. CONTINUE PARCEL ID UNTIL ALL PLOTS MANAGED BY THIS CROP ID PLOT ID RESPONDENT HAVE BEEN COVERED. CROP NAME PID % METHOD 1 1 1 1 1 … 1 1 1 2 1 2 1 2 1 3 1 3 1 3 2 1 1 2 1 … 2 1 2 1 2 1 … 1 Appendix II. Agricultural Reference Questionnaire 73 MODULE 6. CROP ROSTER SEEDS/SEASONAL CROPS 6 7 8 8b 8c 9 10 11 What was Was the main variety traditional Is the variety Did you Were the other varieties of When did What was the When was most of the [CROP] the name of or local? recyclable? use other [CROP] on [PLOT]…? you last buy TOTAL quantity seed planted on [PLOT] the main seed varieties of READ OPTIONS. SELECT ALL the seed you of [CROP] seed during the [REFERENCE variety for this [CROP] on THAT APPLY. planted on planted on [PLOT] AGRICULTURAL SEASON]? MODERN VARIETY, YES.......1 [CROP] on this [PLOT]? this [PLOT]? since the beginning CERTIFIED SEED....................... 1 NO.......2 January............1 plot? of [REFERENCE MODERN VARIETY, MODERN VARIETIES, February.........2 AGRICULTURAL UNCERTIFIED SEED................. 2 YES.......1 CERTIFIED SEED..............................1 March..............3 SEASON]? TRADITIONAL VARIETY, NO.......2 9 MODERN VARIETIES, April................4 UNCERTIFIED SEED................. 3 8b UNCERTIFIED SEED........................2 May..................5 TRADITIONAL VARIETIES, SEE SEED June..................6 UNCERTIFIED SEED........................3 UNIT CODES July...................7 August.............8 September.....9 October.......10 PARCEL ID November....11 CROP ID PLOT ID December....12 15 4-DIGIT YEAR QTY UNIT MONTH YEAR 1 1 1 1 1 … 1 1 1 2 1 2 1 2 1 3 1 3 1 3 2 1 1 2 1 … 2 1 2 1 2 1 … 1 74 AGRICULTURAL SURVEY DESIGN MODULE 6. CROP ROSTER TREE/PERMANENT CROPS EXPECTATIONS 12 13 14 15 16 How many [TREE / In what year were most of these Were the planted [TREE/PERMANENT How much of [CROP] do you expect to Do you intend to sell any of the PERMANENT CROP] trees/ [TREE/PERMANENT CROP] CROP] on [PLOT] trees/plants/ harvest from [PLOT] during [REFERENCE [CROP] that will be harvested plants are on this [PLOT]? trees/plants planted on [PLOT]? seeds…? AGRICULTURAL SEASON]? on [PLOT] during [REFERENCE AGRICULTURAL SEASON] IF DON’T KNOW RECORD READ OUT AND SELECT ALL THAT either in raw or processed form? IF THE FARMER IS UNABLE TO -99 APPLY. QUANTIFY, RECORD -99 SEE CROP QUANTITY CONDITION, AND YES.......1 MODERN VARIETY, UNIT CODES NO.......2 CERTIFIED SEED....................... 1 PARCEL ID MODERN VARIETY, CROP ID PLOT ID UNCERTIFIED SEED................. 2 TRADITIONAL VARIETY, UNCERTIFIED SEED................. 3 NUMBER 4-DIGIT YEAR QTY UNIT CONDITION 1 1 1 1 1 … 1 1 1 2 1 2 1 2 1 3 1 3 1 3 2 1 1 2 1 … 2 1 2 1 2 1 … 1 Appendix II. Agricultural Reference Questionnaire 75 MODULE 7. SEED ACQUISITION LEFTOVER FROM PREVIOUS HARVEST 0 1 2 3 RECORD THE ID OF What is the main reason your household Was any of the In total, how many left THE RESPONDENT. chose to use this [CROP SEED] [REFERENCE [CROP SEED] that over [CROP SEED] LIST FROM AGRICULTURAL SEASON]? your household used did your household ENUMERATOR: LIST ALL SEED TYPES PLANTED FROM since the beginning use [REFERENCE HOUSEHOLD ROSTER SECTION 6 WITH THE ASSOCIATED CROP CODE. RECORD of [REFERENCE AGRICULTURAL IF SEED WAS TRADITIONAL OR IMPROVED FROM MODULE 6, OTHER VARIETIES ARE TOO EXPENSIVE.............1 AGRICULTURAL SEASON]? QUESTION 7 (FIELD CROPS) OR QUESTION 14 (TREE CROPS). OTHER VARIETIES ARE NOT AVAILABLE.............2 SEASON] left over ADVICE FROM EXTENSION OFFICER.................3 from a previous season’s ADVICE FROM INPUT SUPPLIER............................4 harvest? IF BOTH TRADITIONAL AND IMPROVED SEEDS WERE USED ADVICE FROM FELLOW FARMER..........................5 SEE SEED FOR THE SAME CROP, REPORT EACH ON A SEPARATE LINE. HIGHER YIELDS............................................................6 UNIT CODES PREFERS VARIETY........................................................7 SEE SEED STILL HAD SEEDS........................................................8 UNIT CODES ALWAYS USED THIS VARIETY..................................9 OTHER(SPECIFY......................................................-99 YES.......1 NO.......2 4 SEED TYPE CROP CODE PID QTY UNIT (IMPROVED OR TRADITIONAL) 76 AGRICULTURAL SURVEY DESIGN MODULE 7. SEED ACQUISITION FREE SEED PURCHASED SEED 4 5 6 7 8 9 10 Did your household In total, how many free [CROP From whom did you receive How much did Did your household In total, how many purchased Who in your household paid for receive for free SEED] has your household used most of the free [CROP SEED] you pay for purchase any [CROP SEED] has your household the purchased [CROP SEED]? any of the [CROP since the beginning of [REFERENCE that was used [REFERENCE transportation to [CROP SEED] that used since the beginning of PROBE AND SELECT ALL THAT SEED] used since AGRICULTURAL SEASON]? AGRICULTURAL SEASON]? acquire all of the was used since [REFERENCE AGRICULTURAL APPLIES the beginning of [CROP SEED] the beginning of SEASON]? [REFERENCE that you received [REFERENCE LIST UP TO 2: COPY ID FROM HH ONLY RECORD PURCHASED ROSTER” AGRICULTURAL for free during AGRICULTURAL SEEDS USED. EXCLUDE SEEDS SEASON]? SEE SEED [REFERENCE SEASON]? PURCHASED THAT WERE NOT UNIT CODES AGRICULTURAL EXCLUDE ANY USED. [SEED] LEFT OVER SEASON]? YES.......1 FROM PREVIOUS NO.......2 NEXT SEE SEED SEASON UNIT CODES YES.......1 NO.......2 8 QTY UNIT NWID NWID $ QTY UNIT ID CODE ID CODE Appendix II. Agricultural Reference Questionnaire 77 MODULE 7. SEED ACQUISITION PURCHASED SEED 11 12 13 14 15 16 17 Who/which firm/ How much did your What was the value How did you finance this [SEED] How much did you How much did you What was the source of credit for the institution was household pay for of the [CROP SEED] purchase? pay up-front for this repay/will you repay? [SEED] purchase? the main source transport to acquire that your household READ RESPONSES [SEED] purchase? INCLUDE CASH LIST UP TO 2 FROM NETWORK of [CROP SEED] all of the purchased purchased since the INCLUDE CASH PAYMENTS AND ROSTER that you purchased [CROP SEEDS] since beginning of the PAYMENTS AND ESTIMATED VALUE since the beginning the beginning of the agricultural season? PAID IN FULL, WITH OWN-SAVINGS............................... 1 NEXT ESTIMATED VALUE OF IN-KIND of [REFERENCE agricultural season? OF IN-KIND PAYMENTS. AGRICULTURAL RECEIVED ON CREDIT................. 2 INCLUDE ALL TRIPS PAYMENTS. SEASON]? PART OWN-SAVINGS, FROM AND BACK PART ON CREDIT........................... 3 IF NOTHING, TO YOUR FARM. RECORD ZERO. NWID $ $ $ $ $ NWID NWID 78 AGRICULTURAL SURVEY DESIGN NETWORK ROSTER ID 1. NAME 2. CODE 3. LOCATION CODES FOR QUESTION 2: CODES FOR QUESTION 3: RELATIVE.................................................................................1 WITHIN THE VILLAGE........................................................1 NW1 FRIEND/NEIGHBOR............................................................2 NEAR THE VILLAGE.............................................................2 VDC MEMBER........................................................................3 IN/NEAR THE TOWN..........................................................3 NW2 VILLAGE HEADMAN...........................................................4 IN/NEAR THE DISTRICT/URBAN CENTER..................4 TRADITIONAL AUTHORITY............................................5 OUTSIDE THE DISTRICT...................................................5 NW3 POLITICAL LEADER............................................................6 OUTSIDE THE REGION......................................................6 MAIN FARM/PLOT................................................................7 ROADSIDE..............................................................................8 NW4 MOBILE MARKET.................................................................9 LOCAL MARKET................................................................10 NW5 PRIVATE TRADER IN LOCAL MARKET.......................11 LOCAL MERCHANT/GROCERY....................................12 MAIN MARKET....................................................................13 NW6 PRIVATE TRADER IN MAIN MARKET..........................14 AUCTION IN MAIN MARKET........................................15 NW7 PRIVATE COMPANY/BUSINESS PERSON....................16 EMPLOYER............................................................................17 GOVERNMENT AGENCY................................................18 NW8 PARLIAMENT MEMBER.....................................................19 MONEY LENDER/KATAPILA...........................................20 NW9 PRIVATE MICROFINANCE INSTITUTION..................21 SAVINGS & CREDIT COOPERATIVE............................22 NW10 COMMERCIAL BANK.......................................................23 GOVERNMENT-FINANCED LENDER..........................24 PARASTATAL ORGANIZATION.....................................25 NW11 AGRICULTURAL COOPERATIVE..................................26 FARMER BASED CLUB/ASSOCIATION.......................27 NW12 NGO.......................................................................................28 TRUST....................................................................................29 PRIVATE VETERINARY.......................................................30 NW13 DISTRICT VETERINARY....................................................31 RELIGIOUS GROUP/INSTITUTION..............................32 NW14 OTHER (SPECIFY................................................................33 VILLAGE OPEN FORUM...................................................34 NW15 SELF-VACCINNATION......................................................35 Appendix II. Agricultural Reference Questionnaire 79 POST-PLANTING UNIT APPENDIX: EXAMPLE LISTS TO BE CUSTOMIZED FOR CONTEXT CROP CODES SEED UNIT CODES CROP CODE CROP CODE CROP CODE UNIT CODE UNIT CODE BEANS/COWPEA 1010 GARDEN EGG 2080 ZOBO 2290 Kilograms (Kg) 1 Piece Small 80 CASSAVA 1020 GARLIC 2090 APPLE 3010 Grams (g) 2 Piece Medium 81 COCOYAM 1040 GINGER 2100 CASHEW 3020 Litres (l) 3 Piece Large 82 COTTON 1050 OKRO 2120 COCOA 3040 Centilitres (cl) 4 GROUND NUT/PEANUTS 1060 ONION 2130 COCONUT 3050 Heap Small 90 GUINEA CORN/SORGHUM 1070 SWEET/BELL PEPPER (TATASHE) 2141 COFFE 3060 Bin/basket 10 Heap Medium 91 MAIZE 1080 SMALL PEPPER (RODO) 2142 GRAPE FRUIT 3080 Paint Rubber 11 Heap Large 92 MELON/EGUSI 1090 CHILLI PEPPER (SHOMBO) 3030 GUAVA 3090 Milk cup 12 MILLET/MAIWA 1100 PIGEON PEA 2150 KOLANUT 3110 Cigarette cup 13 Bunch Small 100 RICE 1110 PINEAPPLE 2160 LEMON 3120 Tin 14 Bunch Medium 101 WHITE YAM 1121 PLANTAIN 2170 LIME 3130 Bunch Large 102 YELLOW YAM 1122 IRISH POTATO 2180 MANDARIN/TANGERINE 3150 Congo small 20 WATER YAM 1123 SWEET POTATO 2181 MANGO 3160 Congo large 21 Basket Small 140 THREE LEAVE YAM 1124 PUMPKIN 2190 ORANGE 3170 Basket Medium 141 ACHA 2010 GREEN VEGETABLE 2194 OIL PALM TREE 3180 Mudu Small 30 Basket Large 142 BAMBARA NUT 2020 SOYA BEANS 2220 AGBONO(ORO SEED) 3190 Mudu Large 31 BANANA 2030 SUGAR CANE 2230 OIL BEAN 3200 Basin Small 150 BEENI-SEED/SESAME 2040 TEA 2240 PAWPAW 3210 Derica Small 40 Basin Medium 151 CARROT 2050 TOBACCO 2250 PEAR 3220 Derica Medium 41 Basin Large 152 CUCUMBER 2060 TOMATO 2260 AVOCADO PEAR 3221 Derica Large 42 CABBAGE 2070 WALNUT 2270 RUBBER 3230 Derica Very Large 43 Bundle Small 160 LETUS 2071 WHEAT 2280 OTHER(SPECIFY) 9999 Bundle Medium 161 Tiya Small 50 Bundle Large 162 TREE/PERMANENT CROPS Tiya Medium 51 Tiya Large 52 Wheelbarrow 170 Pick-up 180 Kobiowu Small 60 Kobiowu Medium 61 Other (specify) 900 Kobiowu Large 62 Bowl Small 70 Bowl Medium 71 Bowl Large 72 80 AGRICULTURAL SURVEY DESIGN LSMS-ISA REFERENCE QUESTIONNAIRE THIS INFORMATION IS STRICTLY CONFIDENTIAL AND IS TO BE USED FOR STATISTICAL PURPOSES ONLY. POST-HARVEST AGRICULTURE QUESTIONNAIRE HOUSEHOLD IDENTIFICATION CODE NAME 1. DISTRICT: ...................................................................................................................................................................................................................... 2. STATE: ...................................................................................................................................................................................................................... 3. ENUMERATION AREA: 4.VILLAGE NAME: ...................................................................................................................................................................................................................... 5. HOUSEHOLD ID (FROM LIST): 6. NAME OF HOUSEHOLD HEAD: ...................................................................................................................................................................................................................... LATITUDE (S) LONGITUDE (E) 7. GPS COORDINATES OF DWELLING __ __ O __ __ . __ __ __ __ __ __ O __ __ . __ __ __ SURVEY STAFF DETAILS 8. ENUMERATOR CODE: ...................................................................................................................................................................................................................... 9. DATE OF INTERVIEW DD MM YYYY 10. SUPERVISOR CODE: ...................................................................................................................................................................................................................... 11. DATE OF INSPECTION: DD MM YYYY RECORD GENERAL NOTES ABOUT THE INTERVIEW AND ANY SPECIAL INFORMATION THAT WILL BE HELPFUL FOR SUPERVISORS AND DATA ANALYSIS. PLEASE MARK AN `X’ IN BOX IF HOUSEHOLD REFUSAL. PROVIDE DETAILS. Appendix II. Agricultural Reference Questionnaire 81 TABLE OF CONTENTS MODULE 1 Parcel & plot roster MODULE 2 Additional land MODULE 3 Input use MODULE 4 Input roster MODULE 5A Labor inputs (household) MODULE 5B Labor inputs (hired & exchange) MODULE 6 Plot-crop roster MODULE 7 Field crop production MODULE 8 Field crop disposition MODULE 9 Tree & permanent crop production MODULE 10 Tree & permanent crop disposition MODULE 11A Post-harvest labor (household) MODULE 11B Post-harvest labor (hired & exchange) MODULE 12 Farm implements, machinery, and structures MODULE 13 Extension services Network roster Post-harvest unit appendix 82 AGRICULTURAL SURVEY DESIGN MODULE 1. PARCEL & PLOT ROSTER 1 2 3 4 5 6 LIST OF PARCELS FROM THE LIST OF PLOTS FROM POST-PLANTING WAS [PLOT] WAS PLOT Who in the What is the area of this [PLOT]? POST-PLANTING VISIT VISIT LISTED AS MEASURED household was the ENUMERATOR: ASK THE FARMER TO CULTIVATED USING GPS IN primary manager of ESTIMATE THE AREA. DURING THE THE PREVIOUS [PLOT] during PP? POST-PLANTING VISIT? PREFILL VISIT? CODES FOR UNIT: PREFILL ACRE................................ 1 PREFILL HECTARE........................ 2 YES.......1 SQUARE METERS......... 3 YES.......1 NO.......2 PARCEL ID YARDS.............................. 4 NO.......2 OTHER (SPECIFY)......... 5 PLOT ID PARCEL LOCATION PLOT LOCATION & PARCEL NAME PLOT NAME PID AREA UNIT & DESCRIPTION DESCRIPTION 1 1 1 2 1 3 1 4 1 5 2 1 2 2 2 3 2 4 … … Appendix II. Agricultural Reference Questionnaire 83 MODULE 2. ADDITIONAL LAND 1 Is there any other land which your household currently owns or cultivates? YES.......1 NO.......2 NEXT MODULE 2 3 4 5 6 7 LIST ALL ADDITIONAL LIST ALL PLOTS ON What is the area of this What was the primary Who in the What crops were cultivated on [PLOT] PARCELS. ADDITIONAL PARCELS. [PLOT]? use of [PLOT] during household makes during the [REFERENCE AGRICULTURAL RECORD ALL PARCELS ENUMERATOR: ASK the [REFERENCE the decisions SEASON]? BEFORE MOVING TO THE FARMER TO AGRICULTURAL SEASON]? concerning crops INCLUDE FIELD CROPS AND TREE/ QUESTION 3.” ESTIMATE THE AREA. to be planted, PERMANENT CROPS. input use and the CULTIVATED..............................1 LIST UP TO 5 timing of cropping CODES FOR UNIT: RENTED OUT...........................2 activities on this ACRE................................ 1 GIVEN OUT FOR FREE...........3 [PLOT]? HECTARE........................ 2 FALLOW......................................4 FOREST / WOODLOT.............5 THIS DOES SEE CROP CODES SQUARE METERS......... 3 NOT NEED TO YARDS.............................. 4 PASTURE.....................................6 CORRESPOND TO OTHER (SPECIFY)......... 5 SHARECROPPED OUT ..........7 THE OWNER OTHER (SPECIFY).................. 99 PARCEL ID IF NOT CULTIVATED PLOT ID NEXT PLOT PARCEL PLOT PARCEL LOCATION & PLOT NAME LOCATION & AREA UNIT PID CROP 1 CROP 2 CROP 3 CROP 4 CROP 5 NAME DESCRIPTION DESCRIPTION N1 N1 N1 N2 N1 N3 N1 N4 N1 N5 N2 N1 N2 N2 N2 N3 N2 N4 … … 84 AGRICULTURAL SURVEY DESIGN MODULE 3. INPUT USE LIST ALL PLOTS CULTIVATED IN [REFERENCE AGRICULTURAL SEASON] ORGANIC FERTILIZER USE INORGANIC FERTILIZER USE 1 2 3 4 Did your household use any What was the quantity of Did you use any inorganic What was the total quantity of inorganic fertilizer that you used on [PLOT] the REFERENCE organic fertilizer on [PLOT] organic fertilizer that your fertilizer on [PLOT] AGRICULTURAL SEASON? during the [REFERENCE household used on [PLOT] during the [REFERENCE AGRICULTURAL SEASON]? during the [REFERENCE AGRICULTURAL SEASON]? For example, manure, AGRICULTURAL SEASON]? LIST ALL TYPES USED compost, crop waste, etc. YES.......1 CODES FOR TYPE: CODES FOR UNIT: UNIT CODE NO.......2 5 YES.......1 23:21:0+4S/CHITOWE............................. 1 GRAM............................... 1 KILOGRAMS (kg).......... 1 NO.......2 3 DAP............................................................... 2 KILOGRAM..................... 2 GRAMS (g)....................... 2 CAN.............................................................. 3 2 KG BAG....................... 3 UREA............................................................ 4 3 KG BAG....................... 4 RESPONDENT ID D COMPOUND........................................ 5 5 KG BAG....................... 5 OTHER FERTILIZER (SPECIFY)...........99 10 KG BAG..................... 6 50 KG BAG..................... 7 PARCEL ID OTHER (SPECIFY)....... 99 PLOT ID Inorganic Fertilizer #1 Inorganic Fertilizer #2 QTY UNIT TYPE QTY UNIT TYPE QTY UNIT Appendix II. Agricultural Reference Questionnaire 85 MODULE 3. INPUT USE LIST ALL PLOTS CULTIVATED IN [REFERENCE AGRICULTURAL SEASON] HERBICIDE/PESTICIDE USE EQUIPMENT AND MACHINERY USE 5 6 7 8 DID YOU USE What was the total quantity of herbicides/pesticides that you used on [PLOT] the Did you or any member Which equiment/machinery did you or any ANY HERBICIDES/ [REFERENCE AGRICULTURAL SEASON]? of your household use member of the household use? PESTICIDES ON any equipment/ machines [PLOT] DURING on [PLOT] for crop SELECT ALL THAT APPLY LIST ALL TYPES USED THE [REFERENCE maintanence or harvesting AGRICULTURAL during the [REFERENCE MANUALLY OPERATED AND MACHINE SEASON]? AGRICULTURAL SEASON]? POWERED EQUIPMENT: CROP MAINTENANCE SEED/FERTILIZER DRILLS................................ 1 YES.......1 YES.......1 MANURE SPREADERS...................................... 2 NO.......2 7 NO.......2 NEXT PLOT TRANSPLANTERS............................................. 3 THRESHERS......................................................... 4 WINNOWERS.................................................... 5 CODES FOR TYPE: CODES FOR UNIT: SPRAYERS............................................................. 6 INSECTICIDE................................................................7 GRAM..................................................1 DUSTERS.............................................................. 7 HERBICIDE....................................................................8 KILOGRAM........................................2 FERTILIZER BROADCASTERS........................ 8 FUNGICIDE...................................................................9 LITER...................................................8 FUMIGANT................................................................. 10 MILLILITER.........................................9 MACHINE-POWERED EQUIPMENT: OTHER PESTICIDE/HERBICIDE (SPECIFY)....... 11 OTHER (SPECIFY...........................13 CROP HARVESTING MOWERS FOR GRASS CROPS...................... 9 HAY RAKES....................................................... 10 HAY BALERS...................................................... 11 RESPONDENT ID FORAGE HARVESTERS.................................. 12 FORAGE BLOWERS........................................ 13 COMBINE HARVESTERS............................... 14 PARCEL ID CORN PICKERS............................................... 15 PLOT ID OTHER (SPECIFY............................................. 16 Herbicide/pesticide #1 Herbicide/pesticide #2 TYPE QTY UNIT TYPE QTY UNIT 86 AGRICULTURAL SURVEY DESIGN MODULE 4. INPUT ROSTER PRIMARY RESPONDENT FOR MODULE PURCHASED INPUTS 1 2 3 4 5 ENUMERATOR: DID Did your household "In total, how much [INPUT] has Who in your household paid for Who/which firm/ THE HOUSEHOLD purchase any of your household purchased since the purchased [INPUT]? institution was the REPORT THE USE the [INPUT] that the beginning of the [REFERENCE main source of OF [INPUT TYPE] IN was used since AGRICULTURAL SEASON]? [INPUT] that you MODULE 3? the beginning of LIST UP TO 2 purchased since the [REFERENCE the beginning of CODES FOR UNIT: AGRICULTURAL the [REFERENCE YES.......1 GRAM............................... 1 SEASON]? AGRICULTURAL NO.......2 NEXT INPUT KILOGRAM..................... 2 SEASON]? 2 KG BAG....................... 3 YES.......1 3 KG BAG....................... 4 NO.......2 12 5 KG BAG....................... 5 10 KG BAG..................... 6 INPUT CODE 50 KG BAG..................... 7 MILILITER........................ 8 LITER................................ 9 OTHER(SPECIFY)........ 99 INPUT TYPE QTY UNIT PID PID NWID (can be automated in CAPI) 1 ORGANIC FERTILIZER: COMPOST 2 ORGANIC FERTILIZER: MULCH 3 ORGANIC FERTILIZER: BIOFERTILIZERS 4 ORGANIC FERTILIZER: SOLID DUNG 5 ORGANIC FERTILIZER: LIQUID MANURE 6 ORGANIC FERTILIZER: SLURRY 7 INORGANIC FERTILIZER: DAP 8 INORGANIC FERTILIZER : UREA 9 INORGANIC FERTILIZER: CAN 10 INORGANIC FERTILIZER: D COMPOUND 11 PESTICIDES/HERBICIDES: INSECTICIDE 12 PESTICIDES/HERBICIDES : HERBICIDE (SOLID) 13 PESTICIDES/HERBICIDES : HERBICIDE (LIQUID) 13 PESTICIDES/HERBICIDES: FUNGICIDE 14 PESTICIDES/HERBICIDES: FUMIGANT Appendix II. Agricultural Reference Questionnaire 87 MODULE 4. INPUT ROSTER PRIMARY RESPONDENT FOR MODULE PURCHASED INPUTS 6 7 8 9 10 11 How much did What was the How did you finance this [INPUT] purchase? How much did How much did What was the your household pay value of the READ RESPONSES you pay up-front you repay/will source of credit for transport to [INPUT] that for this [INPUT] you repay? for the [INPUT] acquire all of the your household purchase? INCLUDE purchase? purchased [INPUT] purchased since PAID IN FULL, WITH OWN-SAVINGS.......... 1 12 INCLUDE CASH CASH LIST UP TO 2 since the beginning the beginning of RECEIVED ON CREDIT..................................... 2 PAYMENTS AND PAYMENTS FROM NETWORK of the [REFERENCE the [REFERENCE PART OWN-SAVINGS, PART ON CREDIT.... 3 ESTIMATED AND ROSTER AGRICULTURAL AGRICULTURAL VALUE OF ESTIMATED SEASON]? SEASON]? IN-KIND VALUE OF INPUT CODE INCLUDE ALL TRIPS PAYMENTS. IN-KIND FROM AND BACK TO IF NOTHING, PAYMENTS. YOUR FARM. RECORD ZERO. INPUT TYPE $ $ $ $ $ NWID NWID 1 ORGANIC FERTILIZER: COMPOST 2 ORGANIC FERTILIZER: MULCH 3 ORGANIC FERTILIZER: BIOFERTILIZERS 4 ORGANIC FERTILIZER: SOLID DUNG 5 ORGANIC FERTILIZER: LIQUID MANURE 6 ORGANIC FERTILIZER: SLURRY 7 INORGANIC FERTILIZER: DAP 8 INORGANIC FERTILIZER : UREA 9 INORGANIC FERTILIZER: CAN 10 INORGANIC FERTILIZER: D COMPOUND 11 PESTICIDES/HERBICIDES: INSECTICIDE 12 PESTICIDES/HERBICIDES : HERBICIDE (SOLID) 13 PESTICIDES/HERBICIDES : HERBICIDE (LIQUID) 13 PESTICIDES/HERBICIDES: FUNGICIDE 14 PESTICIDES/HERBICIDES: FUMIGANT 88 AGRICULTURAL SURVEY DESIGN MODULE 4. INPUT ROSTER PRIMARY RESPONDENT FOR MODULE: LEFTOVER FROM PREVIOUS SEASON OWN-PRODUCED ORGANIC FERTILIZER 12 13 14 15 Was any of the How much of the [INPUT] that Did you use How much of the ORGANIC [INPUT] that you you used during the [REFERENCE any ORGANIC FERTILIZER was used out of own- used during the AGRICULTURAL SEASON] was left over FERTILIZER out of production or own animals during [REFERENCE from a previous season? own-production or the [REFERENCE AGRICULTURAL AGRICULTURAL own animals during SEASON]? SEASON] left over the [REFERENCE from a previous AGRICULTURAL CODES FOR UNIT: CODES FOR UNIT: season? SEASON]? GRAM............................... 1 GRAM............................... 1 KILOGRAM..................... 2 KILOGRAM..................... 2 YES.......1 YES.......1 2 KG BAG....................... 3 2 KG BAG....................... 3 NO.......2 14 NO.......2 16 3 KG BAG....................... 4 3 KG BAG....................... 4 5 KG BAG....................... 5 5 KG BAG....................... 5 10 KG BAG..................... 6 10 KG BAG..................... 6 50 KG BAG..................... 7 INPUT CODE 50 KG BAG..................... 7 MILILITER........................ 8 MILILITER........................ 8 LITER................................ 9 LITER................................ 9 OTHER(SPECIFY)........ 99 OTHER(SPECIFY)........ 99 INPUT TYPE QTY UNIT QTY UNIT 1 ORGANIC FERTILIZER: COMPOST 2 ORGANIC FERTILIZER: MULCH 3 ORGANIC FERTILIZER: BIOFERTILIZERS 4 ORGANIC FERTILIZER: SOLID DUNG 5 ORGANIC FERTILIZER: LIQUID MANURE 6 ORGANIC FERTILIZER: SLURRY 7 INORGANIC FERTILIZER: DAP 8 INORGANIC FERTILIZER : UREA 9 INORGANIC FERTILIZER: CAN 10 INORGANIC FERTILIZER: D COMPOUND 11 PESTICIDES/HERBICIDES: INSECTICIDE 12 PESTICIDES/HERBICIDES : HERBICIDE (SOLID) 13 PESTICIDES/HERBICIDES : HERBICIDE (LIQUID) 13 PESTICIDES/HERBICIDES: FUNGICIDE 14 PESTICIDES/HERBICIDES: FUMIGANT Appendix II. Agricultural Reference Questionnaire 89 MODULE 4. INPUT ROSTER PRIMARY RESPONDENT FOR MODULE FREE INPUTS 16 17 18 19 20 Did you receive How much of the [INPUT] From whom did you What was the main mode of transportation How much did any of the [INPUT] did you receive in total for receive most of the used to bring back the free [INPUT] that your household pay used during the free during the [REFERENCE free [INPUT] that was your received for free to your household’s for transportation [REFERENCE AGRICULTURAL SEASON]? used the [REFERENCE farm? to acquire all of AGRICULTURAL AGRICULTURAL the [INPUT] that SEASON] for free? SEASON]? you received for CODES FOR UNIT: ON FOOT..................................................................1 free during the GRAM............................... 1 WHEELBARROW....................................................2 [REFERENCE YES.......1 KILOGRAM..................... 2 BICYCLE.....................................................................3 AGRICULTURAL NO.......2 NEXT INPUT 2 KG BAG....................... 3 ANIMAL......................................................................4 SEASON]? 3 KG BAG....................... 4 CART/PUSH CART..................................................5 5 KG BAG....................... 5 MOTORCYCLE.........................................................6 10 KG BAG..................... 6 TRUCK /BUS /MINIBUS.........................................7 INPUT CODE 50 KG BAG..................... 7 BOAT........................................................................... 8 MILILITER........................ 8 OTHER (SPECIFY)....................................................9 LITER................................ 9 NO TRANSPORT/DELIVERED TO FARM........10 OTHER(SPECIFY)........ 99 INPUT TYPE QTY UNIT NWID1 NWID2 $ 1 ORGANIC FERTILIZER: COMPOST 2 ORGANIC FERTILIZER: MULCH 3 ORGANIC FERTILIZER: BIOFERTILIZERS 4 ORGANIC FERTILIZER: SOLID DUNG 5 ORGANIC FERTILIZER: LIQUID MANURE 6 ORGANIC FERTILIZER: SLURRY 7 INORGANIC FERTILIZER: DAP 8 INORGANIC FERTILIZER : UREA 9 INORGANIC FERTILIZER: CAN 10 INORGANIC FERTILIZER: D COMPOUND 11 PESTICIDES/HERBICIDES: INSECTICIDE 12 PESTICIDES/HERBICIDES : HERBICIDE (SOLID) 13 PESTICIDES/HERBICIDES : HERBICIDE (LIQUID) 13 PESTICIDES/HERBICIDES: FUNGICIDE 14 PESTICIDES/HERBICIDES: FUMIGANT 90 AGRICULTURAL SURVEY DESIGN MODULE 5A. LABOR INPUTS (HOUSEHOLD) HOUSEHOLD LABOR 1. (CAPI) 2 3 4 5 6 WAS [PLOT] RESPONDENT ID Has [NAME] worked on On how many days How many hours What activities has [NAME] perfomed on [PLOT] ssince the CULTIVATED [PLOT] for any activity since has [NAME] worked per day on average last interview on [PP INTERVIEW DATE]? Please exclude DURING THE the last interview on [PP on [PLOT]since the did [NAME] work post harvest activities e.g. threshing/shelling, cleaning, etc. [REFERENCE INTERVIEW DATE]? Please last interview on [PP on [PLOT] since the AGRICULTURAL exclude post harvest activities INTERVIEW DATE]? last interview on [PP SELECT ALL THAT APPLY SEASON]? e.g. threshing/shelling, cleaning, INTERVIEW DATE]? etc. LAND PREPARATION................................................1 YES.......1 PLANTING.....................................................................2 INDIVIDUAL ID NO.......2 NEXT PLOT YES.......1 WEEDING......................................................................3 NO.......2 NEXT PERSON-PLOT RIDGING, FERTILIZING, OTHER NON-HARVEST ACTIVITIES.....................4 PARCEL ID HARVESTING................................................................5 PLOT ID SUPERVISION................................................................6 PID DAYS (TOTAL) HOURS PER DAY CODES 1 1 1 1 1 2 1 1 3 1 1 4 1 1 …. 1 2 1 1 2 2 1 2 3 1 2 4 1 …. …. Appendix II. Agricultural Reference Questionnaire 91 MODULE 5B. LABOR INPUTS (HIRED & EXCHANGE) PRIMARY RESPONDENT FOR MODULE: HIRED LABOR 1 2 3 4 5 6 Since the last Since the last Since the last During those Normally, how much Since the last interview on [PP INTERVIEW interview on interview on [PP interview on [PP days when hired did your household DATE], what activities did [PERSON TYPE] perfom [PP INTERVIEW INTERVIEW DATE], INTERVIEW DATE], [PERSON TYPE] pay per day to the on [PLOT]? Please exclude post harvest activities DATE], has your how many [PERSON how many days worked on [PLOT], hired [PERSON] to e.g. threshing/shelling, cleaning, etc. household hired any TYPE] did you or did a typical hired how many hours work on [PLOT]? [PERSON TYPE] to anyone else in your [PERSON] work on per day did a typical work on [PLOT]? household hire to [PLOT]? [PERSON] work? INDICATE THE SELECT ALL THAT APPLY Please exclude post work on [PLOT]? AMOUNT PAID harvest activities e.g. PER PERSON PER LAND PREPARATION................................................1 threshing/shelling, DAY” PLANTING.....................................................................2 cleaning, etc. WEEDING......................................................................3 RIDGING, FERTILIZING, OTHER NON-HARVEST ACTIVITIES.....................4 PARCEL ID YES.......1 HARVESTING................................................................5 PLOT ID NO.......2 7 SUPERVISION................................................................6 PERSON TYPE NUMBER DAYS HOURS PER DAY $ PER DAY CODES 1 1 MEN 1 1 WOMEN 1 1 CHILDREN (UNDER 15) 1 2 MEN 1 2 WOMEN 1 2 CHILDREN (UNDER 15) 1 3 MEN 1 3 WOMEN 1 3 CHILDREN (UNDER 15) 2 …. …. 92 AGRICULTURAL SURVEY DESIGN MODULE 5B. LABOR INPUTS (HIRED & EXCHANGE) PRIMARY RESPONDENT FOR MODULE: EXCHANGE LABOR 7 8 9 10 11 Since the last interview on [PP Since the last Since the last During those days Since the last interview on [PP INTERVIEW DATE], INTERVIEW DATE], have any interview on [PP interview on [PP when [PERSON what activities did these [PERSON TYPE] perfom [PERSON TYPE] from other INTERVIEW DATE], INTERVIEW DATE], TYPE] worked on [PLOT] without pay? Please exclude post harvest households worked on this [PLOT] how many [PERSON how many days did without pay on activities e.g. threshing/shelling, cleaning, etc. free of charge, as exchange labourers TYPE] worked on a typical [PERSON] [PLOT], how many or to assist for nothing in return? [PLOT] without pay? work wihtout pay on hours per day did a SELECT ALL THAT APPLY [PLOT]? typical [PERSON] work? LAND PREPARATION................................................1 YES.......1 NO.......2 NEXT PERSON TYPE-PLOT PLANTING.....................................................................2 WEEDING......................................................................3 RIDGING, FERTILIZING, OTHER NON-HARVEST ACTIVITIES.....................4 PARCEL ID HARVESTING................................................................5 PLOT ID SUPERVISION................................................................6 PERSON TYPE NUMBER DAYS HOURS PER DAY CODES 1 1 MEN 1 1 WOMEN 1 1 CHILDREN (UNDER 15) 1 2 MEN 1 2 WOMEN 1 2 CHILDREN (UNDER 15) 1 3 MEN 1 3 WOMEN 1 3 CHILDREN (UNDER 15) 2 …. …. Appendix II. Agricultural Reference Questionnaire 93 MODULE 6. PLOT-CROP ROSTER 1 2 3 4 PREFILL ALL PLOT-CROP IS THIS CROP ON [PLOT] A [FIELD In the last visit you indicated that Why did your household not harvest any [CROP] from INFORMATION FROM POST- CROP]? you had planted [CROP] on [PLOT]. [PLOT] during the [REFERENCE AGRICULTURAL PLANTING AND MODULE 2 (NEW Did your household harvest any of SEASON]? LAND) the [CROP] during the [REFERENCE AGRICULTURAL SEASON]? LOST CROP DUE TO DROUGHT/LATE ONSET OF RAIN/ERRATIC RAINFALL.......................................................................01 YES.......1 YES.......1 NEXT ROW LOST CROP DUE TO FLOOD................................................................02 SEE CROP CODES NO.......2 NEXT ROW NO.......2 LOST CROP DUE TO OTHER EXTREME NATURAL EVENTS (EXCESSIVE WIND, HAIL, FROST, ETC.) .............................................03 LOST CROP DUE TO PEST......................................................................04 LOST CROP DUE TO VIOLENCE...........................................................05 LOST CROP DUE TO THEFT...................................................................06 DISAGREEMENT ON LAND OWNERSHIP........................................07 UNABLE TO WORK DUE TO SICKNESS.............................................08 NO AVAILABLE LABOR............................................................................09 NOT HARVEST SEASON ........................................................................10 DELAYED/DEFERRED HARVEST............................................................11 PARCEL ID OTHER (SPECIFY).......................................................................................12 CROP ID PLOT ID CROP NAME 94 AGRICULTURAL SURVEY DESIGN MODULE 7. FIELD CROP PRODUCTION 1 2 3 4 When did your household start the Is the area that has been (or will be) Why was the area harvested of [CROP] on [PLOT] less than During the [REFERENCE harvest of [CROP] from [PLOT]? harvested of [CROP] on [PLOT] less the area planted? AGRICULTURAL SEASON], than the area planted? approximately what percent of [PLOT] planted with [CROP] was JANUARY..............01 DROUGHT/LATE ONSET OF RAIN/ERRATIC RAINFALL....... 01 harvested? FEBRUARY............02 YES.......1 FLOOD.................................................................................................... 02 MARCH.................03 NO.......2 5 OTHER EXTREME NATURAL EVENTS APRIL.....................04 (EXCESSIVE WIND, HAIL, FROST, ETC.) ....................................... 03 MAY........................05 PEST.......................................................................................................... 04 JUNE......................06 THEFT...................................................................................................... 05 JULY........................07 UNABLE TO WORK DUE TO SICKNESS....................................... 06 NO AVAILABLE LABOR...................................................................... 07 RESPONDENT ID AUGUST...............08 SEPTEMBER..........09 NOT HARVEST SEASON .................................................................. 08 OCTOBER............10 OTHER (SPECIFY.................................................................................. 09 PARCEL ID NOVEMBER.........11 CROP ID PLOT ID DECEMBER..........12 PERCENT (%) OF MONTH YEAR PLANTED PLOT AREA Appendix II. Agricultural Reference Questionnaire 95 MODULE 7. FIELD CROP PRODUCTION 5 6 7 8 9 How much [CROP] in total did your household harvest from Has your household What date did your household How much more [CROP] does Who in the household [PLOT] during the [REFERENCE AGRICULTURAL SEASON]? completed harvest of complete harvest? your household expect to harvest made or will make LIST UP TO 2 UNIT/CONDITION COMBINATIONS [CROP] from [PLOT]? from [PLOT]? decisions concerning the use of the total JANUARY..............01 YES.......1 harvested [CROP] SEE CROP CONDITION & UNIT CODES FEBRUARY............02 SEE CROP CONDITION & UNIT from [PLOT]? NO.......2 8 MARCH.................03 CODES APRIL.....................04 LIST UP TO 2 MAY........................05 JUNE......................06 JULY........................07 AUGUST...............08 SEPTEMBER..........09 OCTOBER............10 RESPONDENT ID NOVEMBER.........11 DECEMBER..........12 PARCEL ID 9 CROP ID PLOT ID 1st Unit/Condition 2nd Unit/Condition QTY UNIT CONDITION QTY UNIT CONDITION MONTH YEAR QUANTITY UNIT CONDITION PID PID 96 AGRICULTURAL SURVEY DESIGN MODULE 8. FIELD CROP DISPOSITION UNPROCESSED CROP SALES 1 2 3 4 5 LIST ALL RESPONDENT ID Did your How much of the harvested [CROP] was sold in What was the Who/What were the main buyers/outlets HARVESTED household sell unprocessed form? total value of all for your unprocessed [CROP] sales? FIELD CROPS any unprocessed unprocessed FROM SECTION 7. [CROP] since the [CROP] sales? LIST UP TO 2 FROM NETWORK ROSTER harvest? SEE CROP CONDITION & UNIT CODES ESTIMATE THE VALUE OF IN-KIND PAYMENTS. CROP ID YES.......1 NO.......2 11 CROP NAME PID QTY UNIT CONDITION $ NWID NWID Need to be clear about what is considered 'unprocessed' 1 2 3 4 5 6 7 8 9 Appendix II. Agricultural Reference Questionnaire 97 MODULE 8. FIELD CROP DISPOSITION UNPROCESSED CROP SALES 6 7 8 9 10 When was most of the unprocessed Who in your household kept/decided Approximately how many What was the main mode of transportation What was the total [CROP] sold? what to do with earnings from the sale of transactions took place while associated with unprocessed [CROP] sales? cost of transportation unprocessed [CROP]? selling unprocessed [CROP]? READ RESPONSES associated with LIST UP TO 2 FROM HOUSEHOLD unprocessed [CROP] CODES FOR MONTH: sales? ROSTER JANUARY..............01 ON FOOT.......................................................1 FEBRUARY............02 BICYCLE TAXI................................................2 INCLUDE ALL TRIPS MARCH.................03 OWN BICYCLE OR OXCART..................3 FROM AND BACK TO APRIL.....................04 TRUCK / BUS / MINIBUS............................4 THE FARM. MAY........................05 BUYER PICKED UP THE CROP................5 IF NONE, RECORD JUNE......................06 OTHER (SPECIFY).........................................6 ZERO. JULY........................07 AUGUST...............08 SEPTEMBER..........09 OCTOBER............10 CROP ID NOVEMBER.........11 DECEMBER..........12 MONTH YEAR (4-DIGIT) PID PID NUMBER OF TRANSACTIONS $ 1 2 3 4 5 6 7 8 9 98 AGRICULTURAL SURVEY DESIGN MODULE 8. FIELD CROP DISPOSITION OTHER DISPOSITION 11 12 13 14 15 How much of the harvested [CROP] How much of the harvested How much of the harvested How much of the harvested How much of the harvested has been consumed by household [CROP] during the [REFERENCE [CROP] during the [REFERENCE [CROP] during the [REFERENCE [CROP] during the [REFERENCE members? AGRICULTURAL SEASON] was given AGRICULTURAL SEASON] was given AGRICULTURAL SEASON] was used AGRICULTURAL SEASON] was IF NONE, RECORD ZERO out as gifts? out as reimbursements for land, labour, for animal feed? saved for seed? IF NONE, RECORD ZERO or other inputs? INCLUDE THE QUANTITY OF IF NONE, RECORD ZERO IF NONE, RECORD ZERO CROP USED AS ANIMAL FEED DUE TO PEST DAMAGE. IF NONE, RECORD ZERO CROP ID SEE CROP CONDITION & UNIT CODES SEE CROP CONDITION & UNIT CODES SEE CROP CONDITION & UNIT CODES SEE CROP CONDITION & UNIT CODES QTY UNIT CONDITION QTY UNIT CONDITION QTY UNIT CONDITION QTY UNIT CONDITION QTY UNIT CONDITION 1 2 3 4 5 6 7 8 9 Appendix II. Agricultural Reference Questionnaire 99 MODULE 8. FIELD CROP DISPOSITION OTHER DISPOSITION 16 17 18 19 How much of the the harvested [CROP] during the What was the reason for loss? How much of the harvested [CROP] during the How much of the harvested [CROP] during the [REFERENCE AGRICULTURAL SEASON] was lost to READ RESPONSES. LIST UP [REFERENCE AGRICULTURAL SEASON] was [REFERENCE AGRICULTURAL SEASON] was rotting, insects, rodents, theft, etc. in the post-harvest period? TO 2. processed? used for other non specified reasons? EXCLUDE THE QUANTITY OF CROP USED AS ANIMAL IF NONE, RECORD ZERO IF NONE, RECORD ZERO FEED DUE TO PEST DAMAGE. ROTTING...........................................1 FILL IN OPTION 2 (PERCENTAGE) ONLY IF THE FARMER INSECTS.............................................2 CANNOT PROVIDE AN APPROXIMATE QUANTITY FOR RODENTS/PESTS.............................3 OPTION 1. FLOOD...............................................4 SEE CROP CONDITION & UNIT CODES SEE CROP CONDITION & UNIT CODES THEFT.................................................5 IF NONE, RECORD ZERO AND 18 OTHER (SPECIFY)............................6 CROP ID OPTION 1 OPTION 2 QTY UNIT CONDITION % 1ST 2ND QTY UNIT CONDITION QTY UNIT CONDITION 1 2 3 4 5 6 7 8 9 100 AGRICULTURAL SURVEY DESIGN MODULE 8. FIELD CROP DISPOSITION STORAGE 20 21 22 23 24 25 26 27 28 29 Do you have What is your main How much of the harvested What did you do to What What What What What What any of the method of storage for [CROP] during the protect the stored percentage of percentage of percentage of percentage percentage percentage harvested this crop? [REFERENCE AGRICULTURAL [CROP]? [CROP] did [CROP] did [CROP] did of [CROP] of stored of stored [CROP] in READ RESPONSES SEASON] season is being READ RESPONSES. you store to you store to you store to did you store [CROP] do [CROP] do storage now? stored by your household? LIST UP TO 2. consume as sell ? use as seed to render as you expect to you expect to food in your for planting? payment in use as animal use for other UNPROTECTED PILE....... 1 household ? kind? feeding? reasons? HEAPED IN HOUSE.......... 2 SPRAYING....................... 1 YES....1 SEE CROP CONDITION & SMOKING....................... 2 BAGS IN HOUSE................ 3 NO....2 NEXT UNIT CODES HIRED GUARD.............. 3 CHITANDALA IN HOUSE............................ 4 MAGIC (KUTSIRIKA).... 4 DID NOTHING............. 5 CHITANDALA OUTSIDE.............................. 5 OTHER (SPECIFY)......... 6 TRADITIONAL NKHOKWE......................... 6 IMPROVED NKHOKWE......................... 7 METALLIC SILO.................. 8 CROP ID OTHER (SPECIFY).............. 9 QTY UNIT CONDITION 1ST 2ND % % % % % % 1 2 3 4 5 6 7 8 9 Appendix II. Agricultural Reference Questionnaire 101 MODULE 9. TREE & PERMANENT CROP PRODUCTION 1 2 3 4 5 TREE / PERMANENT TREE / PERMANENT RESPONDENT ID Is the [TREE / “How many [TREE / In what year were most How many plants/ trees CROP NAME CROP CODE PERMANENT CROP] PERMANENT CROP] of these plants/ trees were in production in the cultivation in a plantation trees/plants are on this planted? last 12 months? or scattered in the field? [PLOT]? IF CASSAVA, LEAVE BLANK AND6. PARCEL ID PLANTATION................ 1 CROP ID SCATTERED................... 2 IF THE FARMER IS PLOT ID UNABLE TO QUANTIFY, RECORD 999. PID NUMBER YEAR (4-DIGIT) NUMBER 102 AGRICULTURAL SURVEY DESIGN MODULE 9. TREE & PERMANENT CROP PRODUCTION 6 7 8 9 10 What was the last completed production period for the Were there any Why were there losses before the harvest? How much [TREE / Who in the [TREE / PERMANENT CROP]? losses of [TREE READ RESPONSES. LIST UP TO 2. PERMANENT CROP] did household IF THERE IS NO PARTICULAR PRODUCTION PERIOD, USE / PERMANENT you harvest during the last 12 made or will THE LAST 12 MONTHS. CROP] before the months? make decisions last completed IF NOTHING, RECORD ZERO. concerning harvest? DROUGHT/LATE ONSET OF the use of the RAIN/ERRATIC RAINFALL.......................................... 01 total harvested CODES FOR MONTH: FLOOD............................................................................. 02 CODES FOR UNIT: [CROP] from JANUARY............................................1 OTHER EXTREME NATURAL EVENTS KILOGRAM........................................1 [PLOT]? YES....1 FEBRUARY..........................................2 (EXCESSIVE WIND, HAIL, FROST, ETC.).................. 03 50 KG BAG........................................2 LIST UP TO 2 NO....2 9 MARCH...............................................3 PEST................................................................................... 04 90 KG BAG........................................3 APRIL...................................................4 THEFT............................................................................... 05 PAIL (SMALL......................................4 MAY......................................................5 UNABLE TO WORK DUE TO SICKNESS................ 06 PAIL (LARGE)....................................5 JUNE....................................................6 NO AVAILABLE LABOR............................................... 07 NO. 10 PLATE....................................6 JULY......................................................7 NOT HARVEST SEASON ........................................... 08 NO. 12 PLATE....................................7 AUGUST.............................................8 OTHER (SPECIFY).......................................................... 09 BUNCH...............................................8 SEPTEMBER........................................9 PIECE...................................................9 OCTOBER........................................10 BALE..................................................10 NOVEMBER.....................................11 BASKET (DENGU)..........................11 DECEMBER......................................12 OX-CART.........................................12 NO PARTICULAR PERIOD.........99 OTHER(SPECIFY)...........................13 PARCEL ID CROP ID PLOT ID BEGIN END YEAR YEAR MONTH MONTH 1ST 2ND QTY UNIT CONDITION PID PID (4-DIGIT) (4-DIGIT) Appendix II. Agricultural Reference Questionnaire 103 MODULE 10. TREE & PERMANENT CROP DISPOSITION UNPROCESSED CROP SALES 1 2 3 4 LIST ALL HARVESTED RESPONDENT ID Did your household sell How much of the harvested [TREE/ PERMANENT CROP] was sold in What was the total value TREE/PERMANENT CROPS any unprocessed [TREE/ unprocessed form during the last 12 months? of all unprocessed [TREE/ FROM SECTION 9. PERMANENT CROP] PERMANENT CROP] sales during the last 12 months? in the last 12 months? ESTIMATE THE VALUE OF IN-KIND PAYMENTS. CROP ID YES....1 SEE CROP CONDITION & UNIT CODES NO....2 11 CROP NAME PID QTY UNIT CONDITION $ Need to be clear about what is considered 'unprocessed' 1 2 3 4 5 6 7 8 9 104 AGRICULTURAL SURVEY DESIGN MODULE 10. TREE & PERMANENT CROP DISPOSITION UNPROCESSED CROP SALES 5 6 7 8 9 10 Who/What were the When was most of the Who in your household Approximately how many What was the main mode of What was the total cost of main buyers/outlets for unprocessed [TREE/ PERMANENT kept/decided what to do transactions took place while transportation associated with transportation associated your unprocessed [TREE/ CROP] sold? with earnings from the sale selling unprocessed [TREE/ unprocessed [TREE/ PERMANENT with unprocessed [TREE/ PERMANENT CROP] sales? of unprocessed [TREE/ PERMANENT CROP] in the CROP] sales? PERMANENT CROP] sales LIST UP TO 2 FROM PERMANENT CROP]? last 12 months? READ RESPONSES in the last 12 months? CODES FOR MONTH: NETWORK ROSTER JANUARY................ 1 LIST UP TO 2 FROM INLUDE ALL TRIPS FROM FEBRUARY.............. 2 HOUSEHOLD ROSTER AND BACK TO THE FARM. MARCH................... 3 IF NONE, RECORD ZERO. ON FOOT................................................... 1 APRIL....................... 4 BICYCLE TAXI............................................ 2 MAY.......................... 5 OWN BICYCLE OR OXCART.............. 3 JUNE........................ 6 TRUCK / BUS / MINIBUS........................ 4 JULY.......................... 7 BUYER PICKED UP THE CROP............ 5 AUGUST................. 8 OTHER (SPECIFY)..................................... 6 SEPTEMBER............ 9 OCTOBER............10 CROP ID NOVEMBER.........11 DECEMBER..........12 NWID NWID MONTH YEAR (4-DIGIT) PID PID NUMBER OF TRANSACTIONS $ 1 2 3 4 5 6 7 8 9 Appendix II. Agricultural Reference Questionnaire 105 MODULE 10. TREE & PERMANENT CROP DISPOSITION OTHER DISPOSITION 11 12 13 14 15 How much of the [TREE/ How much of the [TREE/ How much of the [TREE/ How much of the [TREE/ How much of the [TREE/ PERMANENT CROP] harvested in the PERMANENT CROP] harvested in the PERMANENT CROP] harvested in PERMANENT CROP] harvested PERMANENT CROP] harvested in last 12 months has been consumed by last 12 months was given out as gifts? the last 12 months was given out as in the last 12 months was used for the last 12 months was saved for household members? IF NONE, RECORD ZERO reimbursements for land, labour, or animal feed? seed? IF NONE, RECORD ZERO other inputs? INCLUDE THE QUANTITY OF IF NONE, RECORD ZERO IF NONE, RECORD ZERO CROP USED AS ANIMAL FEED DUE TO PEST DAMAGE. IF NONE, RECORD ZERO CROP ID SEE CROP CONDITION & UNIT CODES SEE CROP CONDITION & UNIT CODES SEE CROP CONDITION & UNIT CODES SEE CROP CONDITION & UNIT CODES SEE CROP CONDITION & UNIT CODES QTY UNIT CONDITION QTY UNIT CONDITION QTY UNIT CONDITION QTY UNIT CONDITION QTY UNIT CONDITION 1 2 3 4 5 6 7 8 9 106 AGRICULTURAL SURVEY DESIGN MODULE 10. TREE & PERMANENT CROP DISPOSITION OTHER DISPOSITION 16 17 18 19 How much of the the [TREE/ PERMANENT CROP] What was the reason for loss? How much of the harvested [TREE/ How much of the harvested [TREE/ harvested in the last 12 months was lost to rotting, insects, PERMANENT CROP] during the last 12 PERMANENT CROP] during the last 12 rodents, theft, etc. in the post-harvest period? READ RESPONSES. LIST UP TO 2. months was processed? months was used for other non specified reasons? EXCLUDE THE QUANTITY OF CROP USED AS ANIMAL IF NONE, RECORD ZERO ROTTING........................ 1 FEED DUE TO PEST DAMAGE. IF NONE, RECORD ZERO INSECTS.......................... 2 RODENTS/PESTS.......... 3 FILL IN OPTION 2 (PERCENTAGE) ONLY IF THE FARMER FLOOD............................ 4 CANNOT PROVIDE AN APPROXIMATE QUANTITY FOR THEFT.............................. 5 OPTION 1. OTHER (SPECIFY)......... 6 IF NONE, RECORD ZERO AND18 CROP ID OPTION 1 OPTION 2 QTY UNIT CONDITION % 1ST 2ND QTY UNIT CONDITION QTY UNIT CONDITION 1 2 3 4 5 6 7 8 9 Appendix II. Agricultural Reference Questionnaire 107 MODULE 10. TREE & PERMANENT CROP DISPOSITION STORAGE 20 21 22 23 24 25 26 27 28 29 Do you have What is your main How much of the harvested What did you do to What What What What What What any of the method of storage for [TREE/ PERMANENT CROP] protect the stored percentage percentage percentage percentage percentage of percentage of harvested this crop? during the [REFERENCE [TREE/ PERMANENT of [TREE/ of [TREE/ of [TREE/ of [TREE/ stored [TREE/ stored [TREE/ [TREE/ READ RESPONSES AGRICULTURAL SEASON] CROP]? PERMANENT PERMANENT PERMANENT PERMANENT PERMANENT PERMANENT PERMANENT season is being stored by your READ RESPONSES. CROP] did CROP] did CROP] did CROP] did CROP] do CROP] do CROP] in household? LIST UP TO 2. you store to you store to you store to you store you expect to you expect to storage now? UNPROTECTED consume as sell ? use as seed to render as use as animal use for other PILE.................................... 1 food in your for planting? payment in feeding? reasons? HEAPED IN HOUSE..... 2 SPRAYING......................... 1 household ? kind? YES....1 BAGS IN HOUSE........... 3 SMOKING......................... 2 NO....2 NEXT CHITANDALA HIRED GUARD................ 3 IN HOUSE....................... 4 MAGIC (KUTSIRIKA)..... 4 CHITANDALA DID NOTHING............... 5 OUTSIDE......................... 5 OTHER (SPECIFY............ 6 TRADITIONAL NKHOKWE.................... 6 IMPROVED NKHOKWE.................... 7 SEE CROP CONDITION & METALLIC SILO............. 8 UNIT CODES CROP ID OTHER (SPECIFY)......... 9 QTY UNIT CONDITION 1ST 2ND % % % % % % 1 2 3 4 5 6 7 8 9 108 AGRICULTURAL SURVEY DESIGN MODULE 11. PROCESSED CROPS Q1. in the past 12 months, did the household process any of the agricultural crops it produced? YES.....1 Please only include on-farm processing of own produced commodities.” NO....2 SECTION 12A 2 3 4 5 Did you produce [ITEM] To produce [ITEM] did you use this year's Did your household sell How much of [ITEM] was sold in during the past 12 months? production or stock from the previous agricultural any [ITEM] in the past 12 the last 12 months? year? months? YES....1 NO....2 NEXT LINE ONLY THIS YEAR’S PRODUCTION................................1 YES....1 ITEM ONLY PREVIOUS YEAR’S PRODUCTION.....................2 NO....2 8 - MAINLY THIS YEAR’S PRODCTION................................3 CUSTOMIZE LIST TO REFLECT MOST IMPORTANT MAINLY LAST YEAR’S PRODUCTION...........................4 PRODUCTS IN COUNTRY" QTY UNIT Wheat flour Maize flour Rye flour Rice flour Other flour Husked rice Milled rice Polished, glazed, parboiled or converted rice processed or preserved fruit and vegetables Olive oil Soybean oil Palm oil Other vegetable oils wines spirit drinks tobacco products (cigars, chewing tobacco, etc.) others (specify___________) Appendix II. Agricultural Reference Questionnaire 109 MODULE 11. PROCESSED CROPS 6 7 8 9 10 What was the total value Who in your household How much of [ITEM] has How much of [ITEM] was How much of the [ITEM] was of all [ITEM] sales? decided what to do with been consumed by household given out as gifts in the past given out as reimbursements earnings from the sale of members in the past 12 12 months? for land, labour or processed crops? months? other  inputs?                                                              IF NONE, RECORD ZERO ITEM LIST UP TO 2 FROM IF NONE, RECORD ZERO - HOUSEHOLD ROSTER CUSTOMIZE LIST TO REFLECT MOST IMPORTANT PRODUCTS IN COUNTRY" $ PID PID QTY UNIT QTY UNIT QTY UNIT Wheat flour Maize flour Rye flour Rice flour Other flour Husked rice Milled rice Polished, glazed, parboiled or converted rice processed or preserved fruit and vegetables Olive oil Soybean oil Palm oil Other vegetable oils wines spirit drinks tobacco products (cigars, chewing tobacco, etc.) others (specify___________) 110 AGRICULTURAL SURVEY DESIGN MODULE 12A. POST-HARVEST LABOR (HOUSEHOLD) 1. (CAPI) 2 3 4 5 6 WAS [CROP] RESPONDENT ID Since [CROP] was harvested, has Since [CROP] was Since [CROP] was "What activities has [NAME] perfomed on [CROP] since HARVESTED DURING [NAME] worked on any post- harvested, on how harvested, how many your household harvested it? the [REFERENCE harvest activities for [CROP], many days has hours per day did SELECT ALL THAT APPLY AGRICULTURAL beginning with shelling, threshing, [NAME] worked [NAME] typically SEASON]? etc.? on any post-harvest work on [CROP] for INDIVIDUAL ID activities for [CROP]? post-harvest activities? SHELLING/THRESHING/PEELING............................................ 1 CROP CODE DRYING............................................................................................ 2 YES....1 YES....1 CLEANING...................................................................................... 3 NO....2 NEXT CROP NO....2 NEXT PERSON-CROP PROCESSING (MILLING, GRINDING, GRATING, COOKING OIL PRODUCTION, ETC).................................... 4 PID DAYS (TOTAL) HOURS PER DAY CODES 1 1 1 2 1 3 1 4 1 …. 2 1 2 2 2 3 2 4 …. …. Appendix II. Agricultural Reference Questionnaire 111 MODULE 12B. POST-HARVEST LABOR (HIRED & EXCHANGE) PRIMARY RESPONDENT FOR MODULE: HIRED LABOR ENUMERATOR 1 2 3 4 5 6 (CAPI): WAS Since [CROP] Since [CROP] Since [CROP] During those Normally, how Since [CROP] was harvested, what post-harvest [CROP] was harvested, was harvested, was harvested, days when much did your activities did [PERSON TYPE] perfom on HARVESTED has your how many how many days hired [PERSON household pay [CROP]? DURING the household hired [PERSON did a typical TYPE] worked per day to the [REFERENCE any [PERSON TYPE] did your hired [PERSON] on [CROP], how hired [PERSON] AGRICULTURAL TYPE] to work household hire work on many hours per to work on post- SELECT ALL THAT APPLY SEASON]?S on post-harvest to work on post-harvest day did a typical harvest activities activities on post-harvest activities on [PERSON] work? on [CROP]? SHELLING/THRESHING/PEELING...................................1 [CROP]? activities on [CROP]? CROP CODE DRYING...................................................................................2 [CROP]? INDICATE THE AMOUNT PAID CLEANING.............................................................................3 PER PERSON PROCESSING (MILLING, GRINDING, GRATING, COOKING OIL PRODUCTION, ETC)...........................4 PER DAY PERSON TYPE NUMBER DAYS HOURS PER DAY $ PER DAY CODES 1 MEN 1 WOMEN 1 CHILDREN (UNDER 15) 2 MEN 2 WOMEN 2 CHILDREN (UNDER 15) 3 MEN 3 WOMEN 3 CHILDREN (UNDER 15) …. …. 112 AGRICULTURAL SURVEY DESIGN MODULE 12B. POST-HARVEST LABOR (HIRED & EXCHANGE) EXCHANGE LABOR 7 8 9 10 11 Since the [CROP] was harvested, Since [CROP] was Since [CROP] was During those days Since [CROP] was harvested, what post-harvest have any [PERSON TYPE] from harvested, how many harvested, how many when [PERSON TYPE] activities did hired [PERSON TYPE] perfom on [CROP] other households worked on post- [PERSON TYPE] days did a typical worked without pay without pay? harvest activities on [CROP] free worked on post- [PERSON] work on post-harvest of charge, as exchange labourers or harvest activities on wihtout pay on post- activities on [CROP], SELECT ALL THAT APPLY to assist for nothing in return? [CROP] without pay? harvest activities on how many hours [CROP]? per day did a typical SHELLING/THRESHING/PEELING..................................................... 1 [PERSON] work? CROP CODE YES....1 DRYING..................................................................................................... 2 NO....2 NEXT PERSON TYPE-CROP CLEANING............................................................................................... 3 PROCESSING (MILLING, GRINDING, GRATING, COOKING OIL PRODUCTION, ETC)............................................. 4 PERSON TYPE NUMBER DAYS HOURS PER DAY CODES 1 MEN 1 WOMEN 1 CHILDREN (UNDER 15) 2 MEN 2 WOMEN 2 CHILDREN (UNDER 15) 3 MEN 3 WOMEN 3 CHILDREN (UNDER 15) …. …. Appendix II. Agricultural Reference Questionnaire 113 MODULE 13. FARM IMPLEMENTS, MACHINERY, AND STRUCTURES PRIMARY RESPONDENT FOR MODULE: 1 2 4 5 Does your household currently How many [ITEM] does your What is the age of the [ITEM]? If you wanted to sell one of this Did your household use any own [ITEM] ? household currently own? IF MORE THAN ONE [ITEM], [ITEM] today, how much would [ITEM] during the last 12 you receive? months? ITEM CODE ASK FOR THE AVERAGE AGE YES....1 OF ALL [ITEM]S. IF MORE THAN ONE [ITEM], NO....2 5 ASK FOR THE AVERAGE VALUE. YES....1 7 NO....2 ITEM NUMBER YEARS $ IMPLEMENTS 1 HAND HOE 2 SLASHER 3 AXE 4 SPRAYER 5 PANGA KNIFE 6 SICKLE 7 TREADLE PUMP 8 WATERING CAN MACHINERY 9 OX CART 10 OX PLOUGH 11 TRACTOR 12 TRACTOR PLOUGH 13 RIDGER 14 CULTIVATOR 15 GENERATOR 16 MOTORISED PUMP 17 GRAIN MILL 18 OTHER (SPECIFY) STRUCTURES/BUILDINGS 19 CHICKEN HOUSE 20 LIVESTOCK KRAAL 21 POULTRY KRAAL 22 STORAGE HOUSE 23 GRANARY 24 BARN 25 PIG STY 114 AGRICULTURAL SURVEY DESIGN MODULE 13. FARM IMPLEMENTS, MACHINERY, AND STRUCTURES 6 7 8 9 10 11 What was the main reason for Did your household How many [ITEM] did How much did your Did your household rent How much in total did not using the [ITEM]? rent in or borrow any your household rent or household pay to rent or out any [ITEM] during the your household receive [ITEM] during the last borrow during the last borrow [ITEM] during the last 12 months? from the rental of [ITEM] 12 months? 12 months? last 12 months? in the last 12 months? NO NEED FOR ONE.............1 ITEM CODE NEEDS REPAIRS.......................2 ESTIMATE THE VALUE OF YES....1 LENT TO OTHERS..................3 YES....1 IN-KIND PAYMENTS NO....2 NEXT ITEM RENTED TO OTHERS............4 NO....2 10 OTHER (SPECIFY)...................5 ITEM NUMBER $ $ IMPLEMENTS 1 HAND HOE 2 SLASHER 3 AXE 4 SPRAYER 5 PANGA KNIFE 6 SICKLE 7 TREADLE PUMP 8 WATERING CAN MACHINERY 9 OX CART 10 OX PLOUGH 11 TRACTOR 12 TRACTOR PLOUGH 13 RIDGER 14 CULTIVATOR 15 GENERATOR 16 MOTORISED PUMP 17 GRAIN MILL 18 OTHER (SPECIFY) STRUCTURES/BUILDINGS 19 CHICKEN HOUSE 20 LIVESTOCK KRAAL 21 POULTRY KRAAL 22 STORAGE HOUSE 23 GRANARY 24 BARN 25 PIG STY Appendix II. Agricultural Reference Questionnaire 115 MODULE 14. EXTENSION SERVICES PRIMARY RESPONDENT FOR MODULE: 1 2 3 During the [REFERENCE AGRICULTURAL SEASON], did you or Which topic(s) did your household receive information from/ Who in your household received advice/ information through anyone in your household receive any agricultural advice from/ through [SOURCE] on? [SOURCE] during this agricultural (rainy) season? through [SOURCE]? LIST UP TO 4 LIST UP TO 4 ASK FOR EACH TOPIC BEFORE (2). IF NO TO ALL (NEXT SECTION) NEW SEED VARIETIES.........................................................1 PEST CONTROL...................................................................2 FERTILIZER USE.....................................................................3 YES....1 IRRIGATION...........................................................................4 NO....2 COMPOSTING (MANURE.................................................5 MARKETING/CROP SALES................................................6 GROWING/SELLING TOBACCO....................................7 SOURCE CODE ACCESS TO CREDIT............................................................8 FORESTRY...............................................................................9 GENERAL ANIMAL CARE................................................10 ANIMAL DISEASES / VACCINATION............................11 FISHERY PRODUCTION...................................................12 CODE #1 CODE #2 CODE #3 CODE #4 PID #1 PID #2 PID #3 PID #4 1 Government Agricultural Extension Service 2 Private Agricultural Extension Service 3 Government Fishery Extension Service 4 NGO 5 Agricultural Cooperative / Farmers' Association 6 Fishing Cooperative 7 Farmer Field Days / Field School 8 Village Agricultural Extension Meeting 9 Agricultural Extension Course 10 Lead Farmer 11 Peer Farmer (Neighbor / Relative) 12 Electronic Media (TV, Radio, Etc…) 13 Paper Media (Handouts/Flyers) 14 Other (Specify) 116 AGRICULTURAL SURVEY DESIGN MODULE 14. EXTENSION SERVICES PRIMARY RESPONDENT FOR MODULE: 4 5 6 7 8 9 How many times did someone How many times did you How many times did you or Did you or any member of How much in total did your On average, how useful was the from [SOURCE] visit any or any member of your any member of your household your household pay anything household pay to receive advice / information received household member’s farm household visit or meet with attend [SOURCE] in the last in order to receive any type advice / information from from [SOURCE]? during the [REFERENCE the [SOURCE] at a location 12 months? of advice / information from [SOURCE] during the last 12 AGRICULTURAL SEASON]? other than your dwelling or RECORD THE TOTAL FOR [SOURCE] during the last 12 months? VERY USEFUL....................................1 IF NONE, RECORD ZERO your plots during the last 12 THE ENTIRE HH. months? INCLUDE ESTIMATED VALUE SOMEWHAT USEFUL.....................2 months? OF IN-KIND PAYMENTS IF NONE, RECORD ZERO. NOT VERY USEFUL..........................3 SOURCE CODE RECORD THE TOTAL FOR NOT USEFUL....................................4 THE ENTIRE HH. IF NONE, HARMFUL..........................................5 YES....1 RECORD ZERO. NO....2 TOTAL VISITS TOTAL VISITS TOTAL VISITS NUMBER NUMBER NUMBER $ 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Appendix II. Agricultural Reference Questionnaire 117 NETWORK ROSTER ID 1. NAME 2. CODE 3. LOCATION CODES FOR QUESTION 2: CODES FOR QUESTION 3: RELATIVE.................................................................................1 WITHIN THE VILLAGE........................................................1 NW1 FRIEND/NEIGHBOR............................................................2 NEAR THE VILLAGE.............................................................2 VDC MEMBER........................................................................3 IN/NEAR THE TOWN..........................................................3 NW2 VILLAGE HEADMAN...........................................................4 IN/NEAR THE DISTRICT/URBAN CENTER..................4 TRADITIONAL AUTHORITY............................................5 OUTSIDE THE DISTRICT...................................................5 NW3 POLITICAL LEADER............................................................6 OUTSIDE THE REGION......................................................6 MAIN FARM/PLOT................................................................7 ROADSIDE..............................................................................8 NW4 MOBILE MARKET.................................................................9 LOCAL MARKET................................................................10 NW5 PRIVATE TRADER IN LOCAL MARKET.......................11 LOCAL MERCHANT/GROCERY....................................12 MAIN MARKET....................................................................13 NW6 PRIVATE TRADER IN MAIN MARKET..........................14 AUCTION IN MAIN MARKET........................................15 NW7 PRIVATE COMPANY/BUSINESS PERSON....................16 EMPLOYER............................................................................17 GOVERNMENT AGENCY................................................18 NW8 PARLIAMENT MEMBER.....................................................19 MONEY LENDER/KATAPILA...........................................20 NW9 PRIVATE MICROFINANCE INSTITUTION..................21 SAVINGS & CREDIT COOPERATIVE............................22 NW10 COMMERCIAL BANK.......................................................23 GOVERNMENT-FINANCED LENDER..........................24 PARASTATAL ORGANIZATION.....................................25 NW11 AGRICULTURAL COOPERATIVE..................................26 FARMER BASED CLUB/ASSOCIATION.......................27 NW12 NGO.......................................................................................28 TRUST....................................................................................29 PRIVATE VETERINARY.......................................................30 NW13 DISTRICT VETERINARY....................................................31 RELIGIOUS GROUP/INSTITUTION..............................32 NW14 OTHER (SPECIFY................................................................33 VILLAGE OPEN FORUM...................................................34 NW15 SELF-VACCINNATION......................................................35 118 AGRICULTURAL SURVEY DESIGN POST-HARVEST UNIT APPENDIX CROP CODES CROP-SPECIFIC CONDITION CODES CROP PRODUCTION & SALES UNIT CODES TO BE ADDED FROM PP Crop Code Acceptable Conditions TO BE CUSTOMIZED FOR COUNTRY CONTEXT 11 Green harvested – with cob, husk & stalk 12 Green harvested – with cob, husk (no stalk) 13 Green harvested – on the cob (no husk, no stalk) 21 Fresh/raw harvested – with cob & stalk x Maize 22 Fresh/raw harvested – with cob (no stalk) 23 Fresh/raw harvested – on the cob (no husk, no stalk) 31 Dry harvested – with cob, husk & stalk 32 Dry harvested – with cob, husk (no stalk) 33 Dry harvested – on the cob (no husk, no stalk) x Wheat 21 Fresh/raw harvested – with shell & stalk x Barley 22 Fresh/raw harvested – with shell (no stalk) x Rice 31 Dry harvested – with shell & stalk x Finger millet 32 Dry harvested - with shell (no stalk) x Sorghum x Beans 14 Green harvested – in pods or shells x Field peas 24 Fresh/raw harvested – in pods or shell x Cow peas 35 Dry harvested – grain x Pigeon peas x Chick peas x Groundnuts x Soya beans x Simsim x Sunflower 35 Dry harvested – grain x Cotton 35 Dry harvested x Cocoa 24 Fresh/raw harvested - in pods x Vanilla x Tobacco 10 Green harvested - no state needed ALL OTHERS 20 Fresh raw harvested - no state needed Appendix III. Livestock Reference Questionnaire 119 Appendix III. Livestock Reference Questionnaire Agricultural Survey Design Lessons from the LSMS-ISA and Beyond The livestock reference questionnaire is the questionnaire found in: Zezza, A., Pica-Ciamarra, U., Mugera, H. K., Mwisomba, T., & Okello, P. 2016. Measuring the Role of Livestock in the Household Economy: A Guidebook for Designing Household Survey Questionnaires. Washington DC: World Bank. Editable versions of the Agricultural Survey Design reference questionnaires are available for download at: https://www.worldbank.org/en/programs/lsms/publication/AgriculturalSurveyDesign 120 AGRICULTURAL SURVEY DESIGN LIVESTOCK MODULE - CONTENTS SECTION A Livestock Ownership SECTION B1 Changes in Stock over the Past 12 Months: Large and Medium-Size Animals SECTION B2 Changes in Stock over the Past 3 Months: Poultry SECTION C Breeding, Housing, Water, Feeding, and Hired Labor SECTION D Animal Health SECTION E Milk Production (Off-take) SECTION F Egg Production SECTION G Animal Power SECTION H Dung Throughout the module, the questions forming the short module are highlighted in green. Appendix III. Livestock Reference Questionnaire 121 CONFIDENTIAL COUNTRY NAME National Statistical Office NATIONAL LIVING STANDARDS/AGRICULTURAL SURVEY - YEAR 20XX This information is collected under the [NATIONAL STATISTICAL LAW OR SIMILAR] THIS INFORMATION IS STRICTLY CONFIDENTIAL AND IS TO BE USED FOR STATISTICAL PURPOSES ONLY. LIVESTOCK QUESTIONNAIRE HOUSEHOLD IDENTIFICATION CODE NAME 1. REGION: ...................................................................................................................................................................................................................... 2. DISTRICT: ...................................................................................................................................................................................................................... 3. WARD: ...................................................................................................................................................................................................................... 4.VILLAGE/ENUMERATION AREA: ...................................................................................................................................................................................................................... 5. NEIGHBORHOOD NAME: ...................................................................................................................................................................................................................... 6. HOUSEHOLD ID (FROM LIST): 7. NAME OF HOUSEHOLD HEAD: ...................................................................................................................................................................................................................... Note: An editable version of the template is available on the LSMS website, www.worldbank.org/lsms. 122 AGRICULTURAL SURVEY DESIGN NATIONAL LIVING STANDARDS/AGRICULTURAL SURVEY - YEAR 20XX 8. NAME OF INTERVIEWER .................................................................................................................................................................................. OBSERVATIONS ON THE INTERVIEW RECORD GENERAL NOTES ABOUT THE 9. INTERVIEWER ID INTERVIEW AND RECORD ANY SPECIAL INFORMATION THAT WILL BE HELPFUL FOR 10. INTERVIEW START TIME : SUPERVISORS AND THE ANALYSIS OF THIS QUESTIONNAIRE. 11. DATE OF INTERVIEW / / DD MM YYYY 12. NAME OF RESPONDENT .................................................................................................................................................................................. 13. RESPONDENT ID 14. NAME OF SUPERVISOR .................................................................................................................................................................................. 15. SUPERVISOR ID 16. DATE OF QUESTIONNAIRE INSPECTION / / DD MM YYYY 17. NAME OF DATA ENTRY CLERK .................................................................................................................................................................................. 18. DATA ENTRY CLERK ID 19. DATE OF DATA ENTRY / / DD MM YYYY 20. 2ND DATA ENTRY CLERK ID 21. DATE OF 2ND DATA ENTRY / / DD MM YYYY Appendix III. Livestock Reference Questionnaire 123 SECTION A: OWNERSHIP A. OWNERSHIP Did anyone in the household own or keep Did you keep any How many Of these Who in your Do you and / How many Of these Are you any livestock in the past 12 months? [LIVESTOCK TYPE], [LIVESTOCK [LIVESTOCK household is or a member of of the [LIVESTOCK among the irrespective of who TYPE] TYPE], responsible for your household [LIVESTOCK TYPE] owners of the YES.........1 owns [LIVESTOCK does your how many keeping/managing own all of the TYPE] kept how many [LIVESTOCK NO.........0 SKIP THE LIVESTOCK MODULE TYPE] that you keep? household are cross the [LIVESTOCK [LIVESTOCK by your are cross TYPE] currently or exotic TYPE] currently TYPE] currently household or exotic currently keep? breeds? kept by your kept by your are owned breeds? kept by your YES....1 household? household? household? by you or NO....0 a member IF 0  NEXT of your LIVESTOCK NAME Record person ID YES....... 1 8 YES....1 household? NO....0 NO....... 0 CODE LS NAME NOTE: THE OWNER(S) CODE LS TYPE NOTE: LIST UP CAN BE WITHIN AND/ TO 2 IDs FROM LIVESTOCK OR OUTSIDE THE THE HOUSEHOLD HOUSEHOLD ROSTER TYPE LIVESTOCK number number number number NAME 1 2 3 4a 4b 5 6 7 8 Large 1 Bulls 1.0 ruminants Oxen 1.1 Cows 1.2 Steers/Heifers 1.3 Calves - Males/Female 1.4 Buffaloes 1.5 Small 2 Goats - He/She/Kids 2.0 ruminants Sheep - Rams/Ewes/Lambs 2.1 Camelids 3 Camels - He/She/Kids 3.1 Pigs 4 Pigs - Boar/Sow/Piglets 4.1 Poultry 5 Chicken - Cocks / Broilers 5.1 Chicken - Hens / Layers 5.2 Pullets/ DOCs 5.3 Other (Ducks, Geese, Guineafowls, etc.) 5.4 Equines 6 Horses 6.1 Mules / Donkeys 6.2 Other 7 Specify ……………….. 7.0 124 AGRICULTURAL SURVEY DESIGN SECTION A: OWNERSHIP A. OWNERSHIP Who (else) in your Do you or How many Of these Are you Who (else) in your What are the household’s household are among a member of such [LIVESTOCK among the household are among main reasons for owning/ the owners of the of your [LIVESTOCK TYPE], owners of the owners of the keeping [LIVESTOCK [LIVESTOCK TYPE] household TYPE] is how many any of the [LIVESTOCK TYPE] that TYPE]? currently kept by your own any owned by you are cross [LIVESTOCK is not currently kept by household? [LIVESTOCK or a member or exotic TYPE] that your household? sale of live animals..................1 Record person ID TYPE] that of your breeds? is not kept sale of livestock products.....2 is not kept household? by your food for the family.................3 NOTE: LIST UP TO by your household? Record person ID 2 IDs FROM THE savings and insurance............4 household? social status/prestige.............5 HOUSEHOLD ROSTER NOTE: LIST UP TO 2 IDs IF THERE ARE NO YES....1 FROM THE HOUSEHOLD crop agriculture OTHER OWNERS YES....... 1 NO....0 ROSTER. (manure, draught power).....6 WITHIN THE NO....... 0 15 IF THERE ARE NO OTHER transport..................................7 HOUSEHOLD, WRITE OWNERS WITHIN THE other (specify).........................8 CODE = 98. IF THERE HOUSEHOLD, WRITE ARE OWNERS OUTSIDE CODE = 98. IF THERE THE HOUSEHOLD, ARE OWNERS OUTSIDE WRITE CODE = 99. THE THE HOUSEHOLD, CODE LS NAME QUESTION ALLOWS CODE LS TYPE WRITE CODE = 99. THE FOR SIMULTANEOUS QUESTION ALLOWS FOR LIVESTOCK USE OF THE CODES 98 SIMULTANEOUS USE OF AND 99, AS APPLICABLE THE CODES 98 AND 99, AS APPLICABLE TYPE LIVESTOCK number number NAME 9a 9b 10 11 12 13 14a 14b 15 Large 1 Bulls 1.0 ruminants Oxen 1.1 Cows 1.2 Steers/Heifers 1.3 Calves - Males/Female 1.4 Buffaloes 1.5 Small 2 Goats - He/She/Kids 2.0 ruminants Sheep - Rams/Ewes/Lambs 2.1 Camelids 3 Camels - He/She/Kids 3.1 Pigs 4 Pigs - Boar/Sow/Piglets 4.1 Poultry 5 Chicken - Cocks / Broilers 5.1 Chicken - Hens / Layers 5.2 Pullets/ DOCs 5.3 Other (Ducks, Geese, Guineafowls, etc.) 5.4 Equines 6 Horses 6.1 Mules / Donkeys 6.2 Other 7 Specify ……………….. 7.0 Appendix III. Livestock Reference Questionnaire 125 SECTION B: LIVESTOCK STOCK FLAP KEPT [LIVESTOCK NAME] - SEE SECTION A, QUESTION 1 OWNED/KEPT [LIVESTOCK TYPE]? CODE CODE LS LS YES....1 YES....1 LIVESTOCK TYPE TYPE LIVESTOCK NAME NAME NO....0 NO....0 Large ruminants 1 Bulls 1.0 Oxen 1.1 Cows 1.2 Steers/Heifers 1.3 Calves - Males/Female 1.4 Buffaloes 1.5 Small ruminants 2 Goats - He/She/Kids 2.0 Sheep - Rams/Ewes/Lambs 2.1 Camelids 3 Camels - He/She/Kids 3.1 Pigs 4 Pigs - Boar/Sow/Piglets 4.1 Poultry 5 Chicken - Cocks / Broilers 5.1 Chicken - Hens / Layers 5.2 Pullets/ DOCs 5.3 Other (Ducks, Geese, Guineafowls, etc.) 5.4 Equines 6 Horses 6.1 Mules / Donkeys 6.2 Other 7 Specify ……………….. 7.0 126 AGRICULTURAL SURVEY DESIGN SECTION B1: CHANGE IN STOCK - LARGE AND MEDIUM-SIZED ANIMALS BORN PURCHASES GIFTS - RECEIVED & GIVEN CHECK FLAP. IF HOUSEHOLD KEPT ANY [LIVESTOCK NAME] How many How many Has this How many What was Has this How many Has this How many ASK QUESTION 16 IF NOT GO TO NEXT [LIVESTOCK NAME] [LIVESTOCK [LIVESTOCK household [LIVESTOCK the total household [LIVESTOCK household [LIVESTOCK NAME] NAME] purchased NAME] value of the received any NAME] were given any NAME] were did the were born in any live did this [LIVESTOCK [LIVESTOCK received [LIVESTOCK given as a gift household the past 12 [LIVESTOCK household NAME] NAME] as a gift in NAME] in the past 12 keep 12 months? NAME] in buy alive in purchased in as a gift in the past 12 as a gift in months? months ago? the past 12 the past 12 the past 12 the past 12 months? the past 12 months? months months? months? months? CODE LS NAME CODE LS TYPE YES....... 1 YES....... 1 YES....... 1 NO....... 0 21 NO....... 0 23 NO....... 0 25 LIVESTOCK number number number local currency number number TYPE LIVESTOCK NAME 16 17 18 19 20 21 22 23 24 Large 1 Bulls 1.0 ruminants Oxen 1.1 Cows 1.2 Steers/Heifers 1.3 Calves - Males/Female 1.4 Buffaloes 1.5 Small 2 Goats - He/She/Kids 2.0 ruminants Sheep - Rams/Ewes/Lambs 2.1 Camelids 3 Camels - He/She/Kids 3.1 Pigs 4 Pigs - Boar/Sow/Piglets 4.1 Poultry 5 Chicken - Cocks / Broilers 5.1 Chicken - Hens / Layers 5.2 Pullets/ DOCs 5.3 Other (Ducks, Geese, Guineafowls, etc.) 5.4 Equines 6 Horses 6.1 Mules / Donkeys 6.2 Other 7 Specify ……………….. 7.0 Appendix III. Livestock Reference Questionnaire 127 SECTION B1 - CHANGE IN STOCK - LARGE AND MEDIUM-SIZED ANIMALS LOST SALES OF LIVE ANIMALS ANIMALS SLAUGHTERED Has this How many Has this How many What were Has this How many Has this What were household [LIVESTOCK household [LIVESTOCK the total household [LIVESTOCK household the total lost any NAME] sold any NAME] revenues slaughtered NAME] sold some revenues [LIVESTOCK were lost in [LIVESTOCK has this from these any has this of the from these NAME] in the past 12 NAME] alive household [LIVESTOCK [LIVESTOCK household slaughtered sales? the past 12 months? in the past 12 sold alive in NAME] sales? NAME] in slaughtered [LIVESTOCK months? months? the past 12 the past 12 in the past 12 NAME] or (e.g. due months? months? months? their meat in to disease, YES....... 1 the past 12 natural YES....... 1 months? NO....... 0 30 calamity, NO....... 0 injury, theft, IF 0NEXT YES....... 1 etc.) CODE LS NAME LINE NO....... 0 CODE LS TYPE IF 0NEXT YES....... 1 LINE NO....... 0 27 local local number number number LIVESTOCK currency currency TYPE LIVESTOCK NAME 25 26 27 28 29 30 31 32 33 Large 1 Bulls 1.0 ruminants Oxen 1.1 Cows 1.2 Steers/Heifers 1.3 Calves - Males/Female 1.4 Buffaloes 1.5 Small 2 Goats - He/She/Kids 2.0 ruminants Sheep - Rams/Ewes/Lambs 2.1 Camelids 3 Camels - He/She/Kids 3.1 Pigs 4 Pigs - Boar/Sow/Piglets 4.1 Poultry 5 Chicken - Cocks / Broilers 5.1 Chicken - Hens / Layers 5.2 Pullets/ DOCs 5.3 Other (Ducks, Geese, Guineafowls, etc.) 5.4 Equines 6 Horses 6.1 Mules / Donkeys 6.2 Other 7 Specify ……………….. 7.0 128 AGRICULTURAL SURVEY DESIGN SECTION B2: CHANGE IN STOCK - POULTRY BORN PURCHASES GIFTS - RECEIVED & GIVEN CHECK FLAP. IF HOUSEHOLD KEPT ANY [LIVESTOCK NAME] How many How many Has this How many What was Has this How many Has this How many ASK QUESTION 34 IF NOT GO TO NEXT [LIVESTOCK NAME] [LIVESTOCK [LIVESTOCK household [LIVESTOCK the total household [LIVESTOCK household [LIVESTOCK NAME] NAME] purchased NAME] value of the received any NAME] were given any NAME] were did the were born any live did this [LIVESTOCK [LIVESTOCK received [LIVESTOCK given as a gift household in the past 3 [LIVESTOCK household NAME] NAME] as a gift in NAME] in the past 3 keep 3 months? NAME] in buy alive in purchased as a gift in the past 3 as a gift in months? months ago? the past 3 the past 3 in the past 3 the past 3 months? the past 3 CODE LS NAME CODE LS TYPE months? months months? months? months? YES....... 1 YES....... 1 YES....... 1 NO....... 0 39 NO....... 0 41 NO....... 0 43 local number number number number number LIVESTOCK currency TYPE LIVESTOCK NAME 34 35 36 37 38 39 40 41 42 Large 1 Bulls 1.0 ruminants Oxen 1.1 Cows 1.2 Steers/Heifers 1.3 Calves - Males/Female 1.4 Buffaloes 1.5 Small 2 Goats - He/She/Kids 2.0 ruminants Sheep - Rams/Ewes/Lambs 2.1 Camelids 3 Camels - He/She/Kids 3.1 Pigs 4 Pigs - Boar/Sow/Piglets 4.1 Poultry 5 Chicken - Cocks / Broilers 5.1 Chicken - Hens / Layers 5.2 Pullets/ DOCs 5.3 Other (Ducks, Geese, Guineafowls, etc.) 5.4 Equines 6 Horses 6.1 Mules / Donkeys 6.2 Other 7 Specify ……………….. 7.0 Appendix III. Livestock Reference Questionnaire 129 SECTION B2: CHANGE IN STOCK - POULTRY LOST SALES OF LIVE ANIMALS ANIMALS SLAUGHTERED Has this How many Has this How many What were Has this How many Has this What were household [LIVESTOCK household [LIVESTOCK the total household [LIVESTOCK household the total lost any NAME] sold any NAME] revenues slaughtered NAME] sold some revenues [LIVESTOCK were lost in [LIVESTOCK has this from these any has this of the from these NAME] in the past 3 NAME] alive household [LIVESTOCK [LIVESTOCK household slaughtered sales? the past 3 months? in the past 3 sold alive in NAME] sales? NAME] in slaughtered [LIVESTOCK months? months? the past 3 the past 3 in the past 3 NAME] (e.g. due months? months? months? or meat in to disease, the past 3 natural YES....... 1 months? calamity, NO....... 0 48 YES....1 injury, theft, NO....0 CODE LS NAME etc.) IF 0NEXT YES....1 CODE LS TYPE LINE NO....0 YES....... 1 IF 0 NEXT NO....... 0 45 LINE local number number local currency number LIVESTOCK currency TYPE LIVESTOCK NAME 43 44 45 46 47 48 49 50 51 Large 1 Bulls 1.0 ruminants Oxen 1.1 Cows 1.2 Steers/Heifers 1.3 Calves - Males/Female 1.4 Buffaloes 1.5 Small 2 Goats - He/She/Kids 2.0 ruminants Sheep - Rams/Ewes/Lambs 2.1 Camelids 3 Camels - He/She/Kids 3.1 Pigs 4 Pigs - Boar/Sow/Piglets 4.1 Poultry 5 Chicken - Cocks / Broilers 5.1 Chicken - Hens / Layers 5.2 Pullets/ DOCs 5.3 Other (Ducks, Geese, Guineafowls, etc.) 5.4 Equines 6 Horses 6.1 Mules / Donkeys 6.2 Other 7 Specify ……………….. 7.0 130 AGRICULTURAL SURVEY DESIGN SECTION C: BREEDING, HOUSING, WATER, FEED & HIRED LABOR BREEDING HOUSING CHECK FLAP. IF HOUSEHOLD KEPT ANY [LIVESTOCK TYPE] Has this household What has been the main Has this household How much did this What housing system for ASK QUESTION 52 IF NOT GO TO NEXT [LIVESTOCK TYPE] practiced any controlled controlled mating or incurred any cost household spend in [LIVESTOCK TYPE] has this mating or other breeding strategy used related to breeding total over the last 12 household mainly used in the breeding strategy by this household for [LIVESTOCK TYPE]? months for breeding past 12 months? (such as selection of [LIVESTOCK TYPE] in [LIVESTOCK TYPE]? reproductive animal, the past 12 months? YES....... 1 None.................................................1 artificial insemination NO....... 0 56 Confined in sheds..........................2 etc.) for [LIVESTOCK SEE BREEDING CODES Confined in paddocks...................3 TYPE] in the past 12 Confined fences..............................4 months Cage..................................................5 Basket...............................................6 CODE LS NAME CODE LS TYPE YES....... 1 Inside the house (e.g. kitchen) ...7 NO....... 0 56 Other ( specify)..............................8 LIVESTOCK LIVESTOCK local currency TYPE NAME 52 53 54 55 56 Large 1 Bulls 1.0 ruminants Oxen 1.1 Cows 1.2 Steers/Heifers 1.3 Calves - Males/Female 1.4 Buffaloes 1.5 Small 2 Goats - He/She/Kids 2.0 ruminants Sheep - Rams/Ewes/Lambs 2.1 Camelids 3 Camels - He/She/Kids 3.1 Pigs 4 Pigs - Boar/Sow/Piglets 4.1 Poultry 5 Chicken - Cocks / Broilers 5.1 Chicken - Hens / Layers 5.2 Pullets/ DOCs 5.3 Other (Ducks, Geese, Guineafowls, etc.) 5.4 Equines 6 Horses 6.1 Mules / Donkeys 6.2 Other 7 Specify ……………….. 7.0 Appendix III. Livestock Reference Questionnaire 131 SECTION C: BREEDING, HOUSING, WATER, FEED & HIRED LABOR WATERING Who in this household How frequently has What has been the main Has this household How much has this was responsible this household watered source of water for paid for the water for household paid for water for watering the [LIVESTOCK TYPE] in [LIVESTOCK TYPE] in [LIVESTOCK TYPE]? for [LIVESTOCK TYPE] [LIVESTOCK TYPE] in the past 12 month? the past 12 months? the past 12 months? YES....... 1 Record person ID Animals get on their own Borehole...................................1 NO....... 0 61 from available sources...........1 Dam...........................................2 CODE LS NAME Once a day...............................2 Well...........................................3 CODE LSTYPE More than once a day...........3 River/spring/stream................4 Rainwater harvesting.............5 Other (specify).......................6 LIVESTOCK 1st season 2nd season 1st season 2nd season 1st season 2nd season local currency TYPE LIVESTOCK NAME 57 58a 58b 59a 59b 60a 60b 61 Large 1 Bulls 1.0 ruminants Oxen 1.1 Cows 1.2 Steers/Heifers 1.3 Calves - Males/Female 1.4 Buffaloes 1.5 Small 2 Goats - He/She/Kids 2.0 ruminants Sheep - Rams/Ewes/Lambs 2.1 Camelids 3 Camels - He/She/Kids 3.1 Pigs 4 Pigs - Boar/Sow/Piglets 4.1 Poultry 5 Chicken - Cocks / Broilers 5.1 Chicken - Hens / Layers 5.2 Pullets/ DOCs 5.3 Other (Ducks, Geese, Guineafowls, etc.) 5.4 Equines 6 Horses 6.1 Mules / Donkeys 6.2 Other 7 Specify ……………….. 7.0 132 AGRICULTURAL SURVEY DESIGN SECTION C: BREEDING, HOUSING, WATER, FEED & HIRED LABOR FEEDING HIRED LABOR Who in this What have been for this Has this household How much has this Did you hire What was the household was household the major feeding purchased any fodder household spent any labor total cost of the responsible practices for [LIVESTOCK / crop residues / to purchase fodder to help you labor you hired for feeding the TYPE] in the past 12 industrial by-products / crop residues / with livestock for keeping [LIVESTOCK months? / roots & tubers / industrial by-products keeping over livestock over TYPE] in balanced contrates / / roots & tubers / the past 12 the past 12 the past 12 Only grazing/scavenging........1 feed supplements for balanced contrates / months? months? months? [LIVESTOCK TYPE] in feed supplements for Mainly grazing/scavenging, some feeding...........................2 the past 12 months? [LIVESTOCK TYPE] in Mainly feeding, the past 12 months? YES....... 1 YES....... 1 some grazing/scavenging.......3 NO....... 0 NEXT CODE LS NAME Only feeding NO....... 0 66 CODE LS TYPE LINE (zero grazing/scavenging)......4 Other (specify).......................5 local currency person ID 1st season 2nd season 1st season 2nd season local currency LIVESTOCK 1st season 2nd season TYPE LIVESTOCK NAME 62 63a 63b 64a 64b 65a 65b 66 67 Large 1 Bulls 1.0 ruminants Oxen 1.1 Cows 1.2 Steers/Heifers 1.3 Calves - Males/Female 1.4 Buffaloes 1.5 Small 2 Goats - He/She/Kids 2.0 ruminants Sheep - Rams/Ewes/Lambs 2.1 Camelids 3 Camels - He/She/Kids 3.1 Pigs 4 Pigs - Boar/Sow/Piglets 4.1 Poultry 5 Chicken - Cocks / Broilers 5.1 Chicken - Hens / Layers 5.2 Pullets/ DOCs 5.3 Other (Ducks, Geese, Guineafowls, etc.) 5.4 Equines 6 Horses 6.1 Mules / Donkeys 6.2 Other 7 Specify ……………….. 7.0 Appendix III. Livestock Reference Questionnaire 133 SECTION D: ANIMAL HEALTH CHECK FLAP. IF HOUSEHOLD KEPT ANY [LIVESTOCK TYPE] Did [LIVESTOCK TYPE] suffer What kind of disease did Has this household vaccinated Against which diseases are the ASK QUESTION 68 IF NOT GO TO NEXT [LIVESTOCK TYPE] any disease in the past 12 affect [LIVESTOCK TYPE] in any [LIVESTOCK TYPE] in the [LIVESTOCK TYPE] vaccinated? months? the past 12 months? past 12 months? YES....... 1 SEE DISEASE CODES YES....... 1 NO....... 0 70 NO....... 0 72 SEE VACCINATION CODES CODE LS NAME CODE LS TYPE LIVESTOCK TYPE LIVESTOCK NAME 68 69 70 71a 71b 71c 71d Large 1 Bulls 1.0 ruminants Oxen 1.1 Cows 1.2 Steers/Heifers 1.3 Calves - Males/Female 1.4 Buffaloes 1.5 Small 2 Goats - He/She/Kids 2.0 ruminants Sheep - Rams/Ewes/Lambs 2.1 Camelids 3 Camels - He/She/Kids 3.1 Pigs 4 Pigs - Boar/Sow/Piglets 4.1 Poultry 5 Chicken - Cocks / Broilers 5.1 Chicken - Hens / Layers 5.2 Pullets/ DOCs 5.3 Other (Ducks, Geese, Guineafowls, etc.) 5.4 Equines 6 Horses 6.1 Mules / Donkeys 6.2 Other 7 Specify ……………….. 7.0 134 AGRICULTURAL SURVEY DESIGN SECTION D: ANIMAL HEALTH CHECK FLAP. IF HOUSEHOLD KEPT ANY [LIVESTOCK TYPE] During the last 12 During the last 12 During the last 12 During the last 12 During the last 12 ASK QUESTION 68 IF NOT GO TO NEXT [LIVESTOCK TYPE] months has this months has this months how much did months have the months how much did household treated any households treated this household spend on [LIVESTOCK TYPE] in this household spend [LIVESTOCK TYPE] [LIVESTOCK TYPE] vaccines and treatments this household received in total on curative against internal parasites? against external against internal and / any curative treatment? treatments? parasites? external parasites for YES....... 1 [LIVESTOCK TYPE]? YES....... 1 NO....... 0 NEXT LINE YES....1 NO....... 0 NEXT LINE NO....0 CHECK IF QQ. 72,73 CODE LS NAME CODE LS TYPE BOTH ARE = 075 LIVESTOCK local currency local currency TYPE LIVESTOCK NAME 72 73 74 75 76 Large 1 Bulls 1.0 ruminants Oxen 1.1 Cows 1.2 Steers/Heifers 1.3 Calves - Males/Female 1.4 Buffaloes 1.5 Small 2 Goats - He/She/Kids 2.0 ruminants Sheep - Rams/Ewes/Lambs 2.1 Camelids 3 Camels - He/She/Kids 3.1 Pigs 4 Pigs - Boar/Sow/Piglets 4.1 Poultry 5 Chicken - Cocks / Broilers 5.1 Chicken - Hens / Layers 5.2 Pullets/ DOCs 5.3 Other (Ducks, Geese, Guineafowls, etc.) 5.4 Equines 6 Horses 6.1 Mules / Donkeys 6.2 Other 7 Specify ……………….. 7.0 Appendix III. Livestock Reference Questionnaire 135 SECTION E: MILK PRODUCTION (OFF-TAKE) CHECK FLAP. IF HOUSEHOLD KEPT ANY [LIVESTOCK NAME] Did you milk any For how many months During these months During these months Who in your household ASK QUESTION 77 IF NOT GO TO NEXT [LIVESTOCK NAME] [LIVESTOCK TYPE] in on average were in which [LIVESTOCK in which [LIVESTOCK mainly milked the the last 12 months? [LIVESTOCK TYPE] TYPE] were milked, TYPE] were milked, what [LIVESTOCK TYPE] in milked in the past 12 how many animals were was the average quantity the past 12 months ? YES....... 1 months? milked on average each of milk milked per day NO....... 0 NEXT LINE month? from the [LIVESTOCK Record person ID TYPE] herd? CODE LS NAME CODE LS TYPE LIVESTOCK number number litres TYPE LIVESTOCK NAME 77 78 79 80 81 Large 1 Bulls 1.0 ruminants Oxen 1.1 Cows 1.2 Steers/Heifers 1.3 Calves - Males/Female 1.4 Buffaloes 1.5 Small 2 Goats - He/She/Kids 2.0 ruminants Sheep - Rams/Ewes/Lambs 2.1 Camelids 3 Camels - He/She/Kids 3.1 Pigs 4 Pigs - Boar/Sow/Piglets 4.1 Poultry 5 Chicken - Cocks / Broilers 5.1 Chicken - Hens / Layers 5.2 Pullets/ DOCs 5.3 Other (Ducks, Geese, Guineafowls, etc.) 5.4 Equines 6 Horses 6.1 Mules / Donkeys 6.2 Other 7 Specify ……………….. 7.0 136 AGRICULTURAL SURVEY DESIGN SECTION E: MILK PRODUCTION (OFF-TAKE) CHECK FLAP. IF HOUSEHOLD KEPT ANY [LIVESTOCK NAME] What was the main During these months During these months During these months During these months ASK QUESTION 77 IF NOT GO TO NEXT [LIVESTOCK NAME] use of the milk of in which animals were in which animals in which animals were in which animals were [LIVESTOCK TYPE]? milked, how much of were milked, did this milked, how much of the milked, how much did the [LIVESTOCK TYPE] household sell the milk milk of [LIVESTOCK you earn from selling Self consumption....................1 milk collected did this of [LIVESTOCK TYPE]? TYPE] milked did you [LIVESTOCK TYPE] milk Sale............................................2 household consume per sell per week? per week? Processing................................3 week? YES....... 1 CODE LS NAME CODE LS TYPE Other (specify).......................4 NO....... 0 NEXT LINE LIVESTOCK litres litres local currency TYPE LIVESTOCK NAME 82 83 84 85 86 Large 1 Bulls 1.0 ruminants Oxen 1.1 Cows 1.2 Steers/Heifers 1.3 Calves - Males/Female 1.4 Buffaloes 1.5 Small 2 Goats - He/She/Kids 2.0 ruminants Sheep - Rams/Ewes/Lambs 2.1 Camelids 3 Camels - He/She/Kids 3.1 Pigs 4 Pigs - Boar/Sow/Piglets 4.1 Poultry 5 Chicken - Cocks / Broilers 5.1 Chicken - Hens / Layers 5.2 Pullets/ DOCs 5.3 Other (Ducks, Geese, Guineafowls, etc.) 5.4 Equines 6 Horses 6.1 Mules / Donkeys 6.2 Other 7 Specify ……………….. 7.0 Appendix III. Livestock Reference Questionnaire 137 SECTION F: EGG PRODUCTION CHECK FLAP. IF HOUSEHOLD KEPT ANY [LIVESTOCK NAME] Has any How many How many How many Did this How many Who in the How much has ASK QUESTION 87 IF NOT GO TO NEXT [LIVESTOCK NAME] [POULTRY] clutching eggs per [POULTRY] household sell of the hosuehold this household in household periods did clutching did had their the [POULTRY] [POULTRY] did sell the earned produced any [POULTRY] [POULTRY] clutching eggs in the last 3 eggs produced [POULTRY] by selling eggs in the past have on lay on average period in months? did you sell eggs? [POULTRY] CODE LS NAME 12 months? average in in the last the past 3 in the last 3 eggs in the past CODE LS TYPE the last 12 clutching months? YES....... 1 months? Record person 3 months? YES....... 1 months? period? NO....... 0 NEXT ID NO....... 0 NEXT LINE LINE LIVESTOCK number number number number local currency TYPE LIVESTOCK NAME 87 88 89 90 91 92 93 94 Large 1 Bulls 1.0 ruminants Oxen 1.1 Cows 1.2 Steers/Heifers 1.3 Calves - Males/Female 1.4 Buffaloes 1.5 Small 2 Goats - He/She/Kids 2.0 ruminants Sheep - Rams/Ewes/Lambs 2.1 Camelids 3 Camels - He/She/Kids 3.1 Pigs 4 Pigs - Boar/Sow/Piglets 4.1 Poultry 5 Chicken - Cocks / Broilers 5.1 Chicken - Hens / Layers 5.2 Pullets/ DOCs 5.3 Other (Ducks, Geese, Guineafowls, etc.) 5.4 Equines 6 Horses 6.1 Mules / Donkeys 6.2 Other 7 Specify ……………….. 7.0 138 AGRICULTURAL SURVEY DESIGN SECTION G: ANIMAL POWER CHECK FLAP. IF HOUSEHOLD KEPT ANY [LIVESTOCK TYPE] Has this household used any Which have been the main use Has this household used any How much has this household ASK QUESTION 95 IF NOT GO TO NEXT [LIVESTOCK TYPE] of its [LIVESTOCK TYPE] for of [LIVESTOCK TYPE] power in of [LIVESTOCK TYPE] to earned by providing services animal power (e.g. ploughing, the last 12 months provide services (e.g. ploughing, with [LIVESTOCK TYPE] in the transport) past 12 months? transport) to other households? past 12 months? YES....... 1 YES....... 1 Transport.......................................... 1 Crop agriculture NO....... 0 NEXT LINE NO....... 0 NEXT LINE (ploughing, seeding, weeding, CODE LS NAME threshing, milling)............................. 2 CODE LS TYPE Other (specify)................................ 3 LIVESTOCK 1st season 2nd season local currency TYPE LIVESTOCK NAME 95 96a 96b 97 98 Large 1 Bulls 1.0 ruminants Oxen 1.1 Cows 1.2 Steers/Heifers 1.3 Calves - Males/Female 1.4 Buffaloes 1.5 Small 2 Goats - He/She/Kids 2.0 ruminants Sheep - Rams/Ewes/Lambs 2.1 Camelids 3 Camels - He/She/Kids 3.1 Pigs 4 Pigs - Boar/Sow/Piglets 4.1 Poultry 5 Chicken - cocks / broilers 5.1 Chicken - hens / layers 5.2 Pullets/DOCs 5.3 Other (Ducks, Geese, Guineafowls, etc.) 5.4 Equines 6 Horses 6.1 Mules / Donkeys 6.2 Other 7 Specify ……………….. 7.0 Appendix III. Livestock Reference Questionnaire 139 SECTION H: ANIMAL DUNG CHECK FLAP. IF HOUSEHOLD KEPT ANY [LIVESTOCK TYPE] Has this household Which have been the main use of the dung Did this household sell How much has this ASK QUESTION 99 IF NOT GO TO NEXT [LIVESTOCK TYPE] made any use of the produced by [LIVESTOCK TYPE] in the last the dung produced by household earned from the dung produced by 12 months? [LIVESTOCK TYPE] in sales of dung produced by [LIVESTOCK TYPE] in the last 12 months? [LIVESTOCK TYPE] in the the past 12 months? past 12 months? Manure (grazing on the plot or collected and applied)............................................1 YES....... 1 Fuel............................................................................. 2 NO....... 0 NEXT ANIMAL Construction material............................................3 CODE LS NAME Feed to other animals............................................4 CODE LS TYPE YES....... 1 Sale............................................................................. 5 NO....... 0 NEXT ANIMAL Other (specify)........................................................6 LIVESTOCK local currency TYPE LIVESTOCK NAME 99 100a 100b 101 102 Large 1 Bulls 1.0 ruminants Oxen 1.1 Cows 1.2 Steers/Heifers 1.3 Calves - Males/Female 1.4 Buffaloes 1.5 Small 2 Goats - He/She/Kids 2.0 ruminants Sheep - Rams/Ewes/Lambs 2.1 Camelids 3 Camels - He/She/Kids 3.1 Pigs 4 Pigs - Boar/Sow/Piglets 4.1 Poultry 5 Chicken - Cocks / Broilers 5.1 Chicken - Hens / Layers 5.2 Pullets/ DOCs 5.3 Other (Ducks, Geese, Guineafowls, etc.) 5.4 Equines 6 Horses 6.1 Mules / Donkeys 6.2 Other 7 Specify ……………….. 7.0 140 AGRICULTURAL SURVEY DESIGN CODES FOR SECTIONS C AND D BREEDING CODES FOR Q 53 CODES FOR Q69 (Diseases) CODES FOR Q71 (Vaccinations) Natural Mating, Sire Selected from within the herd.................... 1 Brucelosis.................................................... 1 Brucelosis.................................................... 1 Natural Mating, Sire Purchased or exchanged................................ 2 CBPP............................................................ 2 CBPP............................................................ 2 Natural Mating, Non-Breeding Males Castrated........................... 3 Lumpy Skin Disease.................................. 3 Lumpy Skin Disease.................................. 3 Natural Mating, Dam Purchased............................................................ 4 CCPP........................................................... 4 CCPP........................................................... 4 Artificial Insemination................................................................................ 5 ECF............................................................... 5 ECF............................................................... 5 Other (Specify).............................................................................................. 6 Rabies........................................................... 6 Rabies........................................................... 6 FMD............................................................. 7 FMD............................................................. 7 Anthrax....................................................... 8 Anthrax....................................................... 8 Black Quarter (BQ).................................. 9 BQ................................................................ 9 New castle Disease................................10 New castle Disease................................10 Small Pox...................................................11 Small Pox...................................................11 Gomboro..................................................12 Gomboro..................................................12 Helminthiosis...........................................13 Other, specify...........................................13 ASF.............................................................13 Tick Borne Disease................................15 Typanosomiasis........................................16 SELECT LSMS GUIDEBOOKS Disability Measurement in Household Surveys: A Guidebook for Designing Household Survey Questionnaires Marco Tiberti and Valentina Costa January 2020 Trees on Farms: Measuring Their Contribution to Household Welfare Daniel C. Miller, Juan Carlos Muñoz-Mora, Alberta Zezza, and Josefine Durazo September 2019 Food Data Collection in Household Consumption and Expenditure Surveys Prepared by The Inter-Agency and Expert Group on Food Security, Agricultural and Rural Statistics April 2019 Measuring Household Expenditure on Education Gbemisola Oseni, Friedrich Huebler, Kevin McGee, Akuffo Amankwah, Elise Legault, and Andonirina Rakotonarivo December 2018 Spectral Soil Analysis & Household Surveys Sydney Gourlay, Ermias Aynekulu, Calogero Carletto, and Keith Shepherd October 2017 The Use of Non-Standard Units for the Collection of Food Quantity Gbemisola Oseni, Josefine Durazo, and Kevin McGee July 2017 Measuring the Role of Livestock in the Household Economy Alberto Zezza, Ugo Pica-Ciamarra, Harriet K. Mugera, Titus Mwisomba, and Patrick Okell November 2016 Land Area Measurement in Household Surveys Gero Carletto, Sydney Gourlay, Siobhan Murray, and Alberto Zezza August 2016 Measuring Asset Ownership from a Gender Perspective Talip Kilic and Heather Moylan April 2016 Measuring Conflict Exposure in Micro-Level Surveys Tilman Brück, Patricia Justino, Philip Verwimp, and Andrew Tedesco August 2013 Improving the Measurement and Policy Relevance of Migration Information in Multi-topic Household Surveys Alan de Brauw and Calogero Carletto May 2012 Agricultural Household Adaptation to Climate Change: Land Management & Investment Nancy McCarthy December 2011 Living Standards Measurement Study www.worldbank.org/lsms data.worldbank.org