Communication Networks and Nutrition-sensitive Extension in Rural Kenya: Essays on Centrality, Network Effects and Technology Adoption
by Lisa Jäckering
Date of Examination:2018-05-07
Date of issue:2018-05-29
Advisor:Prof. Dr. Meike Wollni
Referee:Prof. Dr. Matin Qaim
Referee:Prof. Dr. Stephan von Cramon-Taubadel
Files in this item
Name:Dissertation_JaeckeringLisa_FINAL.pdf
Size:2.57Mb
Format:PDF
Abstract
English
Globally, 767 million people live on less than US$ 1.90 a day and two billion people are malnourished. Especially affected by poverty and malnutrition is the rural population of Sub-Saharan Africa (SSA), who depend on the agricultural sector for food and income. Adopting new technologies can help farmers improve their livelihoods through an increase in income, or an improved nutritional and health status. However, adoption rates are comparably low. As agriculture can play a central role for food security, making agriculture more nutrition-sensitive has become one of the hot topics in the recent development discourse. However, also the uptake of pro-nutrition technologies – such as biofortified crops or particularly nutritious pulses – remains below expectations. While factors influencing the adoption of technologies are manifold (for instance, education, risk preferences or wealth), special attention has recently been paid to the important functions of information access and social networks. In this regards, agricultural extension systems can set in to provide farmers with the missing information on new (pro-nutrition) technologies. A common approach is to channel information regarding the new technologies through farmer groups. However, so far nutrition-sensitive programs mostly focused on mothers only. There is little evidence on how men and women embedded in groups, communicate about topics related to agriculture and nutrition, and which persons can serve as potential target points for nutrition-sensitive extension. Simultaneously, networks play an important role for the diffusion of information. In particular, communication networks are potential pathways that may induce behavioral change and may play a strong role in the setting of group-based extension due to dynamics that trigger peer pressure or competition. However, due to lack of detailed (panel) network data, there is little evidence on how these communication networks are affected by the delivery of agricultural extension, and if communication networks can contribute to finally adopt new technologies. This dissertation addresses these research gaps by drawing conclusion based on a unique dataset that combines a randomized controlled trial (RCT) with detailed panel data on communication networks of farmer groups. The RCT was implemented in rural Kenya and consisted of varying combinations of group-based agricultural and nutrition training sessions. The purpose of the extension training was the promotion of the iron-rich black common bean variety KK15. Survey data from 48 farmer groups (824 households) was collected before (October until December 2015) and after (October until December 2016) the intervention (March until September 2016). Given the background on the importance of a better understanding of communication networks in the context of agricultural extension, this dissertation comprises two essays. The first essay (Chapter 2) of this dissertation deals with nutrition and agricultural communication networks of farmer groups and builds on baseline data of 48 farmer groups (815 individuals), we collected in 2015: In developing countries, community-based organizations (CBOs) and individuals within CBOs are important target units for agricultural programs. However, little is known about the flow of information within CBOs and between individuals. The objective of this study is to investigate the structure and characteristics of communication networks for nutrition and agriculture. First, we identify the structure of agricultural and nutrition information networks within CBOs, as well as overlaps of the two networks. Dyadic regression techniques are then used to explore the characteristics of persons forming links for agriculture and nutrition. Second, key persons within CBOs that are prominent or influential for agriculture and nutrition information networks are identified, as well as characteristics of persons that are excluded from these networks. Analysis is conducted using descriptive and econometric techniques such as fixed effect Poisson models. Our study finds that nutrition information is exchanged within CBOs but to a moderate extent. Further, agricultural and nutrition information networks overlap and often the same links are used for both topics. At the same time, a large number of people are excluded from nutrition information networks. These persons are more likely to be men, have smaller land sizes and are less connected to persons outside of the group. We conclude that there is room for nutrition training to sensitize group members and nudge communication exchange about nutrition related issues. In particular, we recommend incentivizing communication with isolated persons. Further, our regression results suggest targeting CBO leaders, as well as other group members that live in central locations as an entry point for training. The results can help to increase the outreach of nutrition-sensitive programs. The second essay (Chapter 3) investigates if interventions, such as agricultural extension, affect agricultural communication networks and if these communication networks can act as pathways leading to the adoption of new technologies. The analysis is based on the mentioned RCT and therefore uses both, baseline, as well as follow-up data: A growing body of literature focuses on the role of network effects for farmers’ adoption decisions. However, little is known on how interventions affect networks. We analyze the effect of group-based trainings on networks and the influence of these networks on the adoption of technologies. Our analysis builds on a unique dataset that combines a randomized controlled trial (RCT) with detailed panel data on communication networks. Results suggest that, first, the intervention had a positive impact on communication among farmers (i.e. the creation of communication links). Second, besides positive direct effects of the intervention, we also find strong positive network effects on adoption, indicating that individual farmers are more likely to adopt, the higher the share of adopters in their communication network. Hence, group-based extension approaches can be efficient in diffusing new technologies, not only because they reduce transaction costs, but also because network effects can stimulate and drive technology adoption.
Keywords: Communication networks; Centrality; Community-based organizations; Nutrition-sensitive agriculture; Dyadic regressions; Network effects; Randomized controlled trial; Group-based extension