Least Squares Regression Analysis of Poverty Indicators in Rural Communities of Plateau State

Rural poverty in Plateau State, North Central Nigeria, is shaped by a complex interplay of agricultural productivity, access to social services, household demographics, and environmental vulnerability. This study applies ordinary least squares regression and stepwise variable selection to identify the most significant determinants of poverty incidence across 18 selected rural communities in Mangu, Shendam, and Langtang North Local Government Areas. Primary data are collected from 360 households using a multistage random sampling framework, capturing household income, landholding size, years of education, dependency ratio, distance to nearest market, and access to extension services. The poverty line is defined using the Plateau State-specific household expenditure threshold computed from National Bureau of Statistics consumption data. Model diagnostic tests confirm absence of multicollinearity and heteroscedasticity in the final specification. Results indicate that years of education, landholding size, and market access distance are the three strongest statistically significant predictors, jointly explaining 61 percent of variance in household poverty scores. The marginal effect of an additional year of education corresponds to a 7.4 percent reduction in predicted poverty probability. The study recommends coordinated investment in rural feeder roads and primary school completion programmes as high-impact poverty reduction strategies in the state. Keywords: least squares regression, rural poverty, Plateau State, determinants, household survey.