Multivariate Statistical Analysis of Agricultural Input Use and Productivity Among Smallholder Farmers in Nassarawa State

Understanding the multivariate statistical relationships between agricultural input combinations and productivity outcomes among smallholder farmers in Nassarawa State, North Central Nigeria, is essential for designing effective input subsidy programmes and agricultural extension interventions. This study applies multivariate statistical methods to analyse input use patterns and their productivity implications for 380 smallholder maize and cassava farmers in Lafia, Akwanga, and Doma LGAs of Nassarawa State. Survey data covered fertiliser use, improved seed adoption, pesticide application, irrigation access, labour input, and yield per hectare. Cluster analysis grouped farmers by input use profile. Multivariate analysis of variance tested yield differences between input use clusters. Canonical correlation analysis examined the overall relationship between input use and productivity outcome vectors. Cluster analysis identified four input use profiles: low-input subsistence (42 percent), fertiliser-only (26 percent), improved seed and fertiliser (22 percent), and high-input commercial (10 percent). MANOVA confirmed significant multivariate yield differences between clusters (Wilks Lambda = 0.43, F = 12.8, p < 0.001). Canonical correlation revealed a first canonical variate (Rc = 0.79) linking fertiliser and improved seed use to both high yield and high net returns. The high-input cluster achieved 2.8 times the yield of the low-input cluster. The study provides evidence for the productivity complementarity of fertiliser and improved seed inputs and recommends integrated input packages as the optimal subsidy design for Nassarawa State smallholders. Keywords: multivariate analysis, agricultural inputs, productivity, Nassarawa State, cluster analysis