UBC Theses and Dissertations
UBC Theses and Dissertations
Vancouver personal travel market : a multivariate statistical analysis Nichols, Peter Royall
This thesis has two objectives. The first is to determine the critical variables in the Vancouver personal travel market. The second objective is to suggest additional surveys which might be undertaken. To determine the critical variables in the Vancouver personal travel market, an analysis is made of the data collected during the 1966 Air Canada Vancouver Survey. This analysis involves two separate but mutually reinforcing approaches. The first is concerned with the propensity to travel by air and by all modes combined for any given adult with known characteristics. The second approach is concerned with the evaluation of particular factors affecting the choice between modes of travel. Trips by air are compared to trips by auto, rail and bus in terms of the characteristics of the trips and in terms of the characteristics of the persons taking the trips. In suggesting additional surveys, this study focuses not only on the Air Canada Vancouver Survey but also on other recent travel surveys and analyses by transportation companies, government departments, travel associations, and academic institutions. This part of the analysis involves two main areas of consideration. First, suggestions are made as to additional types of data which might prove valuable in the transport firm's marketing decisions. Second, suggestions are made as to more useful and revealing methods of analyzing the data collected. The study finds past travel market research to be characterized by two main shortcomings. The first is the failure of many studies to consider and analyze the interrelationships between trip variables and market variables. It is suggested that only by matching trip characteristics to market characteristics can the firm get some idea of the type of service, advertising appeal and pricing policy that are most effective for particular segments of the market. The second major shortcoming found in most travel market research is the failure to separate the direct relationships between two variables from their indirect relationships with other variables. While it is recognized that variables such as income and occupation are closely interrelated, very few studies have made any attempt to determine the separate importance of occupation beyond that which is indirectly attributable to income. In the light of the shortcomings found in past travel market research, the analysis of the Vancouver personal travel market involves the use of multivariate statistical methods. The basic method employed is multiple least squares regression. However, the inherent quantitative nature of this statistical procedure necessitates that several modifications be made for its application to the predominantly qualitative market and trip variables in the survey. The primary modification is concerned with the determination of numerical scales for the dependent and independent variables. A second major modification is concerned with the presentation of the multiple regression results. This modification involves taking the mean of the dependent variable as a first approximation of the magnitude to be estimated and adjusting this value according to whether the particular observation being considered belongs to a particular group or possesses a particular characteristic. The thesis concludes that certain market variables and certain trip variables have particularly strong and reliable relationships with personal travel. Income is found to be the single most important market variable, with the domicile of the adult's children and parents being second in significance and reliability. However, trip variables are found to have generally more significant and reliable associations with travel than market variables. The trip variables showing greatest importance are trip distance and trip purpose. The study further concludes that additional surveys of the Vancouver personal travel market should involve deeper investigation of those variables found in this study to be of particular significance and that some attempt should be made not only to filter out the indirect influences between the variables but also to measure the magnitude of these indirect influences.
Item Citations and Data