DEVELOPMENT OF PREDICTIVE MODEL FOR MOTORCYCLE OWNERSHIP IN MAKURDI METROPOLIS, NIGERIA

Abstract

This research work involves development of a predictive model for improving motorcycle ownership and operations in Makurdi metropolis of Benue state. Data on motorcycle ownership and operations were collected from households and motorcycle parks using well structured and closed ended questionnaires aimed at identifying factors affecting motorcycle operations in the study area. A total of 1412 respondents for households and 486 commercial motorcyclists were interviewed by means of the questionnaires. Data collected from the questionnaires was analysed using statistical package for social sciences (SPSS Version 22) and a predictive Multinomial Logit model (MNL), developed that shows the influence of a function that determines outcome probabilities, which functions to estimate the various factors identified via the responses on the motorcycle operations. The result of the descriptive statistical data analysis shows 15%, 29% and 56% of respondents were from low, medium and high density areas respectively and use their motorcycles for private (42%) and commercial (49%) purposes, the greater number of which has at least a post-primary education (88%) and the number of motorcycles owned decreased as the educational qualification and income increased. The models developed for motorcycle ownership for low, medium and high density areas proved adequate, predicting motorcycle ownership of 67%, 63% and 79% respectively and all giving a high R-square value>0.5. The aggregate model for predicting motorcycle ownership gives and estimated ownership of 67% which is close to the 71% obtained from observed questionnaire data and R-square value of >0.5 which makes it adequate. Thus, it is recommended that the developed models can be used in predicting motorcycle