Vol 5 Issue 1 April 2017-September 2017
Jean-Claude Nzabonimpa, Dr. Joseph K. Mung’atu, Dr. Marcel Ndengo
Abstract: The purpose of this research was mainly to measure the level of association between various differentials and household agricultural income in Rwanda. Basing on the nature of the variables in this study, an ordered logistic regression model (OLR) was used to statistically measure the relationship between an ordinal dependent variable (Agricultural income) and a set of independent variables. In analysis the dependent variable was classified into 3 groups namely: Low, Medium and High agricultural income. The discussed household agricultural income in this study were mainly generated from the sales of crops, processed products, livestock and livestock products, auto-consumption, rent of land and agricultural equipments. The Fisher test was used to analyze the goodness of fit of the model and hence it indicated that the model was statistically significant as a whole, meaning that the observed data perfectly corresponded to the fitted model. The findings from this study revealed that households (HHs) owning at least one cow have more chances of being either in the middle or in the highest categories of agricultural income. The results also reiterated the fact that HH size has a significant influence on agricultural income of the households, since HHs with large size were more likely to be found in the lower category of agricultural income. Additionally, the data highlighted that the likelihood for a HH to be either in the middle or in the highest categories of agricultural income rises when that household resides in a rural area.
Keywords: Household Agricultural income Computation, F-test, Treatment of outliers, Imputation, OLR Model.
Title: Measuring the Level of Association between Various Differentials and Household Agricultural Income in Rwanda (2013-2014) using Ordered Logistic Regression Model
Author: Jean-Claude Nzabonimpa, Dr. Joseph K. Mung’atu, Dr. Marcel Ndengo
International Journal of Mathematics and Physical Sciences Research
ISSN 2348-5736 (Online)
Research Publish Journals
