Comparative Impacts of Digital Extension Tools on Smallholder Yields

Thesis Abstract

This study investigates the differential effects of digital extension tools on smallholder yields within a cross-sectional framework addressing persistent gaps in agricultural knowledge transfer and productivity in smallholder systems. The core problem is that digital extension platforms vary in reach, content relevance, and user engagement, yet robust comparative evidence on their relative impact on agronomic outcomes remains limited, hindering targeted policy and program design. The aim is to compare the impacts of three digital extension modalities—SMS advisory services, smartphone-based agro-mobile applications, and interactive voice response (IVR) platforms—on maize and bean yields, input use efficiency, and adoption of recommended practices among smallholder farmers. The study pursues three objectives (1) to estimate the average treatment effects of each digital tool on yield and input efficiency; (2) to identify channels through which tools influence outcomes, including knowledge acquisition, behavioral change, and access to markets; and (3) to examine heterogeneity in effects by farmer characteristics (education, farm size, access to mobile internet, and gender). The methodological design is a cross-sectional, mixed-methods approach. A representative sample of 1,200 smallholder farming households across three agro-ecological zones will be surveyed, stratified by tool exposure, with 400 households per tool group and 200 non-exposed controls. Data collection will combine structured questionnaires capturing socio-economic characteristics, access to digital tools, input use, farming practices, and crop yields, with agronomic records where available. In-depth interviews and focus group discussions (n=40 interviews, 12 focus groups) will explore user experiences, perceived benefits, barriers to adoption, and contextual factors shaping tool effectiveness. Instrument validity will be ensured through pre-testing, expert evaluation, and cognitive interviewing; reliability will be assessed using test-retest methods and Cronbach’s alpha for multi-item scales. Quantitative analysis will employ propensity score matching to address selection bias across exposure groups, followed by augmented inverse probability weighting to estimate average treatment effects on yields and input efficiency. Regression analysis will utilize multivariate linear models controlling for plot-level and household covariates, and a difference-in-differences approach will be explored where longitudinal data or recall-based proxies permit. Mediation analysis will assess pathways through knowledge gains, changes in management practices, and market access. Subgroup analyses will test for heterogeneous effects by education, gender, and access to internet connectivity. Qualitative data will be analyzed thematically using a framework approach, with coding guided by the Technology Acceptance Model and the Diffusion of Innovations theory to triangulate quantitative findings and illuminate implementation nuances. The study will integrate the Sustainable Agricultural Intensification framework to interpret practice adoption and yields in the context of resource constraints and risk. Expected findings include (1) differential yield gains attributable to smartphone-based tools being highest among agronomically informed farmers, with significant but smaller effects for SMS advisory and IVR platforms; (2) greater improvements in input efficiency and adoption rates among users with higher digital literacy and reliable mobile internet access; (3) evidence of knowledge transfer and recommended practice adherence as primary channels driving yield improvements, with market linkages contributing to income effects; and (4) notable gender and education-based disparities in tool benefits, moderated by extension service integration and local content relevance. The study contributes to knowledge by providing rigorous, comparator-based evidence on the relative effectiveness of distinct digital extension modalities for smallholders, informing design and deployment of digital agricultural extension programs, and offering a nuanced understanding of contextual determinants of impact. Policy recommendations will emphasize targeted tool selection aligned with farmer capabilities, investments in digital literacy, and integration of digital tools with conventional extension services to maximize yield gains and sustainable practice adoption. The main conclusion anticipates that digital extension tools generate measurable yield and efficiency gains, with smartphone-based applications offering the strongest average impact when accompanied by supportive literacy and infrastructure, while SMS and IVR platforms remain viable for low-literacy contexts with established content relevancy. Recommendations include prioritizing mixed-tool strategies, expanding affordable data-enabled access, and embedding monitoring and evaluation components to continuously tailor content and delivery mechanisms to farmer needs.