This landscape study provides an overview of Tanzania’s agriculture and the status of country’s extension system. It also provides information on the prevalence of poverty, nutrition and gender related issues in the country with special focus on rural areas. It summarizes Tanzania’s current agricultural and nutrition policy, and provides a summary of several on-going projects by the United States Government (USG) and other donors in the country related to agriculture extension, and gender and nutrition impacts.

Sub-Saharan Africa remains the world’s most food-insecure region in spite of its abundant agricultural potential. In an attempt to contribute towards overcoming this problem, an agricultural development approach known as RIPAT (Rural Initiatives for Participatory Agricultural Transformation) has been developed over the period since 2006 through a series of projects in northern Tanzania. 

It has for decades been anticipated by development actors that pro-poor agricultural development interventions would be the direct route to improved nutrition among smallholder farm families. However, it is difficult to find evidence that documents such linkage – partly because of poor quality evaluations, but also because it has been realised that agricultural development interventions must be designed to a much larger extent with a nutritional lens and must take into account what types of agricultural component can lead to improved nutrition. We provide research evidence of improved rural food and nutrition security following the application of the RIPAT approach. 

The three self-contained chapters of this dissertation evolves around different aspects of a Farmer Field School intervention taking place in northern Tanzania, studying the diffusion of agricultural technologies and how agriculture links to food security, poverty and child health. The intervention is called RIPAT (Rural Initiatives for Participatory Agricultural Transformation) and was funded by the Rockwool Foundation. As a consultant for the Rockwool Foundation I administered a large scale data collection for an impact evaluation of RIPAT.

All three chapters of my dissertation build on these data. The first two chapters are coauthored with Helene Bie Lilleør. A recurrent theme in this dissertation is the identification of causal effects. Since participation in RIPAT is voluntary, the data does not offer direct experimental variation which I can exploit for identification, and there exists no baseline data collected before the implementation of RIPAT I to control for selection. I pursue different identification strategies which are detailed in the three chapters. In the following, I provide a preview of the findings.