Impact Evaluation of Precision Agriculture Adoption in Smallholder Maize Systems

Thesis Abstract

Smallholder maize production in sub-Saharan Africa faces productivity and efficiency constraints underscored by limited access to timely, site-specific agronomic advice. Precision agriculture (PA) offers potential to optimize input use, improve yields, and reduce environmental footprints, yet evidence from smallholder contexts remains fragmented due to heterogeneity in adoption, farm conditions, and extension support. This study aims to evaluate the impact of adopting precision agriculture technologies on productivity, input-use efficiency, and income among smallholder maize farmers. The specific objectives are (i) to quantify the effect of PA adoption on maize yield and gross margin; (ii) to assess changes in per-hectare input use and input cost structure; (iii) to identify determinants of adoption and the role of extension services, farm size, and access to credit; (iv) to analyze farmers’ perceptions of PA’s operational feasibility, reliability, and maintenance requirements; and (v) to examine gender-differentiated and household-level implications of PA adoption. The study adopts a quasi-experimental design employing a matched-parm approach to estimate treatment effects of PA adoption. The population comprises smallholder maize farmers in three agro-ecological zones within a national agricultural transformation program. A sample of 600 households is selected through stratified random sampling, with 300 adopters and 300 non-adopters, matched on farm size, soil fertility, age of household head, input intensity, and prior extension contact. Data are collected via structured household surveys, on-farm crop measurements, and key informant interviews with extension agents and input suppliers across two cropping seasons. Instruments include a standardized questionnaire for household demographics, farm practices, input use, and income; a GPS-enabled yield measurement protocol; and a semi-structured interview guide for qualitative insights. Validity is enhanced through pre-testing, translation-back translation, and pilot testing; reliability is assessed with Cronbach’s alpha for multi-item scales and test-retest for longitudinal measures. Analytical methods combine econometric and qualitative techniques. The primary quantitative model is a propensity score matching (PSM) framework to balance observable characteristics, followed by difference-in-differences (DiD) estimation to capture treatment effects over time. Regression analyses control for endogeneity via instrumental variable approaches where appropriate; robustness checks include placebo tests and alternative matching schemes. Outcome variables include yield (kg/ha), gross margin (local currency/ha), input-use intensity (kg/ha, currency terms), and return on investment. Subgroup analyses explore heterogeneity by farm size, gender of the household head, and access to credit. The qualitative component employs thematic analysis of in-depth interviews to contextualize quantitative findings, with coding derived from the Technology Acceptance Model (TAM) and the Diffusion of Innovations (DOI) framework to interpret adoption determinants, perceived ease of use, perceived usefulness, and organizational constraints. Expected findings anticipate that PA adoption will be associated with statistically significant improvements in yield and gross margins, primarily through enhanced input-use efficiency, precise fertilizer and seed placement, and timely irrigation scheduling where applicable. It is expected that higher adoption will correlate with reduced per-unit input costs and improved risk management under variable rainfall scenarios. Heterogeneity analyses may reveal greater gains for medium-sized farms and male-headed households, contingent on access to extension and credit. The study also expects nuanced perceptions where technical challenges, maintenance costs, and reliability concerns influence sustained use, while perceived agronomic benefits reinforce continued adoption among progressive farmers. The study contributes to knowledge by providing robust, context-specific evidence on the productivity, economic, and behavioral impacts of PA in smallholder maize systems, informing policy design for scalable, farmer-centered digital agriculture interventions. It also advances methodological integration of PSM-DiD with qualitative insights grounded in TAM and DOI, offering a replicable framework for impact assessment of precision farming in similar rural economies. Based on the findings, recommendations will address design of subsidized PA service delivery, capacity-building for maintenance and data interpretation, alignment of credit and input supply chains with PA requirements, and targeted extension strategies to bridge gender and household equity gaps, ensuring that the benefits of precision agriculture translate into sustainable increases in productivity and livelihoods.