ABSTRACT
The journey to work had been the subject of several studies. Results from dome of these studies have been used to formulate models for use in travel demand forecasts. However, virtually all these studies have been carried out in the developed countries. Because of differences in environmental factors, one cannot arbitrarily adapt their findings for use in developing countries like Nigeria. Hence, it is necessary for similar studies to be carried out in Nigeria, so as to evaluate the prospects of transfering results and also to determine the limitations of the study techniques as regards their applicability to local conditions. This work examines the commuting patterns of some randomly selected Civil Servants working in Kaduna State Government Secretariate. The questionaire methods of data collection was utilized in gathering information on the travel behaviour of the selected subjects. Questions asked cover both level of service attributes of the modes used and socio-economic characteristics of the respondents. Discriminant analysis was then used to isolate the major factors responsible for differences in mode choice. Prior to administering the questionaires, the distrubution by income of the total population of Civil Servants, from which samples were taken, was obtained. Proportionate numbers of worker were then randomly chosen within each income group. However, disproportionately small response rate was obtained from the low income groups. This suggests that special attention has to be paid to low income earners when conducting such studies, with the use of questionaires, in order to have a fair representation of the lower income group. Personal income was found to be a great determinar t of the mode selection of the studied workers. It was also found that transport fare weighs heavily in the mode choice decisions of the workers. This letter findings would seem to be at variance to the general belief that travellers, in developing countries, are mode-captive. However, no other level of service attribute was found to be statistically significant from the discriminant analysis. The results sugges that when questionaires are used for data collection in developing countries, special attention need be paid to the low income earners. This would yield a more proportionately uniform data representation across the income groups. The results also suggest that variables like travel times do not contribute significantly to the modechoice decisions, in contrast to findings in similar studies in developer countries. These findings indicate that methods of model development derived in developed countries are not generally transferable to the developing countries.