Determinants of Finger Millet Adoption and Dis-Adoption among Smallholder Farmers in Nakuru, Kenya

Abstract

Kenya's national annual finger millet production is projected at 0.91 tons per hectare for the 2024/2025, significantly below its potential yield of 2.5 tons per hectare. This production gap exists despite the crop’s valuable attributes, including a short cropping cycle, low reliance on external inputs, high nutritional value, and natural resilience to climate shocks, positioning it as a key strategy for improving food security. Between 2014 and 2023, finger millet production in Kenya experienced a severe decline, with a 31.1% drop in total tonnage, a trend driven by both a reduction in cultivated acreage and lower crop yields. This study addresses the paradox that while finger millet adoption is a promising strategy, its viable production is consistently hampered by low adoption, dis-adoption and non-adoption. This phenomenon is hypothesized to be influenced by a complex interplay of socioeconomic, institutional, and psychosocial factors, a topic that has not been extensively documented, particularly in the study area. This research was conducted in Nakuru County with three main objectives: to characterize the finger millet production systems, to assess farmers' knowledge and attitudes toward the crop, and to evaluate the factors determining adoption and dis-adoption. A mixed-methods, cross-sectional household survey design was employed, collecting data from 384 households (adopters, dis-adopters, and non-adopters) identified through a multi-stage purposive sampling process. The research instruments, including structured questionnaires and interviews, demonstrated high reliability with a Cronbach's alpha of 0.82. Data were analyzed using descriptive statistics for characterization, ANOVA to assess group differences, and a multinomial logit (MNL) regression model to determine the key factors influencing adoption decisions. Findings revealed significant demographic and perceptual differences among the farmer groups. Adopters were notably older, with an average age of 51 years, and had higher annual household incomes (KES 504,300) compared to non-adopters. A significant majority of adopting households (62.8%) were also headed by women. The analysis of perceptual factors showed a distinct knowledge and attitudinal gap, with adopters demonstrating a significantly higher knowledge score (81.4%) and more positive attitudes (61.2%) towards the crop. The MNL model further confirmed the critical role of these factors, showing that farmer knowledge and attitude had a significant positive influence on adoption and a negative influence on dis-adoption. In a key finding, the study established that the decision to adopt and sustain production is a dynamic process shaped by a dual effect of resources. The model showed that an increase in household labor, for instance, significantly increased the probability of adoption (ME=0.123, p=0.015) while simultaneously decreasing the probability of dis-adoption (ME=-0.150, p=0.002). Similarly, access to extension services had a significant positive influence on adoption and a negative influence on dis-adoption. In conclusion, this study establishes that sustainable finger millet production is a function of a powerful synergy between a farmer's attributes, particularly knowledge and attitude, and the sustained presence of critical socioeconomic and institutional support. Based on these findings, it is recommended that stakeholders prioritize farmer empowerment through enhanced extension services and farmer groups. Additionally, providing strategic resources like modern mechanization and high-quality seed varieties, alongside gender-inclusive programs specifically supporting female-headed households, is crucial for facilitating sustained adoption.