Effect of Agricultural Extension Education on Adoption of Climate-Smart Maize Technologies: A Biostatistical Analysis of Smallholder Farmers of Chirundu District of Zambia

Abstract

The context is that rain-fed smallholder agriculture is highly vulnerable to climate change in Agro-Ecological Region I (AER I) of Zambia, which is dominated by the low-lying semi-arid valley of Chirundu District in Zambia, part of the most vulnerable districts in the country. Climate-smart maize technologies (CSMTs) have been identified as an important tool for advancing the adoption of climate smart technologies, but there is a lack of empirical evidence to quantify the biostatistical impact of agricultural extension education in AER I settings. This study used robust biostatistical methods, accounting for selection bias and unobservable heterogeneity, to explore the impact of agricultural extension education on the uptake of climate-smart maize technologies among smallholder farmers in Chirundu District, Zambia. Cross sectional survey design was used. A multi-stage stratified random sampling was used to select 320 smallholder farmers in four agricultural camps of Chirundu District. Structured household interviews were conducted from June to August 2024. Evaluated technologies were drought-tolerant maize varieties (DTMVs), conservation agriculture (CA), micro-dose fertilisation and water harvesting technologies. The main biostatistical method used was binary logistic regression, which was used to model the probability of CSMT adoption, with Propensity Score Matching (PSM) used to estimate the Average Treatment Effect on the Treated (ATT) for extension-educated farmers, and chi-square tests conducted for bivariate associations. 58.4% of the 320 respondents had had one or more contacts to agricultural extension education during the last twelve months. Extension education was found to be significant and enhance the chance of CSMT adoption (OR = 3.74; 95% CI: 2.41–5.81; p < 0.001) from binary logistic regression. Years of formal education (OR = 1.28; p = 0.003), household land size (OR = 1.45; p = 0.011), and access to credit (OR = 2.17; p = 0.002) were also significant predictors. Results of PSM analysis showed that the extension-educated farmers were 31 percentage points more likely to adopt CSMTs than matched non-participants, with the ATT being significant at p < 0.001. The chi-square test revealed significant relationships among the extension attended and adoption of DTMVs, conservation agriculture and water harvesting (χ² = 34.72; p < 0.001), (χ² = 29.18; p < 0.001) and (χ² = 22.47; p < 0.001) respectively. The conclusion is that agricultural extension education is a statistically significant and practically meaningful factor to influence the adoption of climate-smart maize technologies by smallholder farmers in Chirundu District. Policy investment on improving extension services, especially in AER I districts that are drought-prone, is critical to increase climate resilience and food security in Zambia.