A Comparison of Probit Regression and Binary Regression
Robinson Amos Ibuchi; Okeregwu Blessing Amaka; Ockiya, Atto Kennedy; & Inamete Emem Ndah H.
Department of Mathematics & Statistics
Ignatius Ajuru University of Education, Rivers State, Nigeria
Email: ibuchirobinsonamos@gmail.com
ABSTRACT
In this research, an evaluation of the relationship between a response variable and several explanatory variables were considered using Binary and probit regression. The methods used in the analysis were descriptive statistics and regression techniques. This research focuses on the household utilized/non utilizes primary health care services with a formulated questionnaire, which were administered to 400 households. The statistical Software packages used are M icrosoft Excel, SPSS 21 and Minitab 16. The result showed that the Binary regression model is the best fit in modelling binary response variable in form of a count data; based on the two assessment criteria employed [Akaike Information Criterions (AIC) and Bayesian Information Criterions (BIC)].
Keywords: Probit regression, Binary regression, Household Utilized/Non Utilizes Primary Health Care Services.
