There are approximately 500 million smallholder farmer households worldwide, at least half of which live in poverty and food insecurity. Scientific research underpins development efforts by providing options for improved varieties, breeds, or practices (termed “interventions”); and by providing analyses of how to increase the adoption and impact of those interventions. One of the most widely used method of data collection to evaluate or predict the impact of interventions is the household survey, but critical evaluations of the effectiveness of household survey data and methodologies are rare. Lack of standardised questions make efforts to aggregate findings across datasets challenging, given that different surveys often yield widely different data, both in terms of content and quality, which severely limits the comparability of those data (see Chapters 2 and 3). Here I present an improved survey method to assess farm practices and food security for smallholder households in lower income countries, primarily tropical or sub-tropical. The tool is named the Rural Household Multi-Indicator Survey (RHoMIS). It makes use of recent advances in
digital technologies, which enables quicker data collection and reporting than in previous generations of survey tools. The tool was designed to be rapid, lean, user-friendly, flexible and reliable (Chapter 3). The design ethic and advances in indicator formulation allowed data to be gathered on a wider range of topics over shorter time frames but still with adequate depth to permit effective analyses (Chapters 4, 5, and 6). During development RHoMIS was deployed by 13
organisations in 17 countries, with over 15,000 interviews conducted. The tool has the flexibility needed for application in many locations, sufficient standardisation to permit rapid analysis and data aggregation between sites, and enables more efficient characterisation of smallholder farming systems compared to previous efforts. Findings of analyses presented in this thesis stress the need to understand the heterogeneity of smallholders, and to plan or evaluate interventions for specific subsets of households. Analyses presented in the research chapters show that the farm strategy of input intensification is better suited to larger farms, crop diversification is better suited to smaller farms, and that the effects are strongly influenced by the degree of commercialisation of livelihoods (Chapter 3), the use of collected resources can strongly benefit the poorest households (Chapter 4), and that off-farm incomes in combination with farm intensification hold the potential to raise the prosperity of about 90% of the households studied in Chapter 5 . Furthermore households show different levels of interest in trialling and adopting new practices which are not necessarily related to their assets or farm types, in Chapter 2 about one quarter of households were identified as likely to trial new practices. In
particular the analyses highlights that those experiencing (or at risk of) extreme food insecurity benefited most from opportunities for off-farm income, whereas moderately poor households benefited more from agricultural intensification. These findings indicate that for agricultural intensification measures to raise households out of food insecurity and poverty they must be targeted to the appropriate group of smallholders, and to succeed must be in combination with
opportunities to earn off farm income.