Discriminant Function Analysis of Factors affecting Off-farm Diversification among Small-scale Farmers in North Central Nigeria

The study evaluated the factors that affected off-farm diversification among small-scale farmers in North Central region of Nigeria. Multistage sampling technique was used to select 180 respondents. The primary data obtained with the aid of standard questionnaire were analysed using discriminant function analysis. The dependent variable, off-farm work typology, comprised three groups namely, agricultural wage employment, non-agricultural wage employment, and self employment. Based on factor loading, the strongest predictor was fund for farm investment (0.654) while the weakest predictor was crop failure (0.359). The canonical correlation of 0.572 implied that 32.72% of the variation in off-farm work typology was explained by the discriminators included in the model. The chi-square statistic (77.89) of Wilk’s lambda was statistically significant (p<0.01), implying that discriminant function was appropriate and significant. Self-employment category had the best classification (88.3%), while agricultural wage had the poorest classification (0.0%). Although, the F-statistic of the Box’s M test (7.07) was significant and the null hypothesis accepted that the covariance matrices were not equivalent, the significance was disregarded on the grounds that the sample size was large and the number of groups in the dependent variable was more than two. It was concluded that small-scale farmers in the study area embarked on enterprise diversification in order to generate funds for farm investment, although there was a gradual drift from the core agricultural production sector to self-employment. Therefore, government should encourage farm investment by subsidising farm input, improving farm produce price and farm credit delivery. This would, among other things, check the adverse impact of dual farm structure on food production. Keywords: Nigeria, small-scale farmers, diversification, off-farm work typology, discriminant function analysis, predictor