An Investigation of the McMaster Commuter Distribution

Abstract

<p> The purpose of this study is to investigate the McMaster commuter distribution. The total analysis is carried out in two subanalyses. The first analysis examines the effect that different variables have on the mean trip length of various commuter groups. The variables examined in the first analysis are: residential tenure, occupation, duration of service or study, parking permit ownership and part-time versus full-time status of students. In general the first analysis is concerned with the mean trip length of on campus Faculty, Staff, and Students. The first analysis verified that residential tenure is an important variable in that students who commute from the homes of parents average a considerably greater mean trip length than students who commute from rented accommodation.</p> <p> The second analysis employs a disaggregate singly-constrained spatial interaction model to distribute trips between McMaster and student residential locations. The second analysis shows that: (i) the production-constrained model fits considerably better with observed data when the sample is partitioned into student renter and stay at home groups than when the sample is not partitioned. The attractiveness factors were varied between the two groups. Renters were considered to be attracted to renter occupied dwellings in a zone while students commuting from the homes of parents were considered to be attracted to the number of owner occupied dwellings in a zone. (ii) straight line distance as a surrogate for travel cost yields a better fit for the renter group while automobile travel-time facilitates a better fit for the stay at home group. Auto travel time yields a better fit for peripheral trips because of the tendency for these trips to be made by car. Given the understanding that student renters are predominantly bus users who have chosen to locate close to the campus, euclidean distance is apparently more reflective of the travel impedance experienced by this group. Future research should attempt to qualitatively link measures of travel cost with the client group they are attempting to model.</p>