In 1984, a survey was designed to determine taxi supply and demand in the Montreal urban area. The main problem was that virtually nothing is known about the taxi driver population due to lack of good records. Secondly, the only records were the taxi companies' radio records. The third problem concerned ensuring that there were a good representation of the trips in both time and space. The sampling method involved having an observer meet with a taxi driver during his shift to give him a questionnaire for the trips he made during the two hours following the initial interview. To select the taxicab, the observer first called the taxi company, pretending to be an ordinary customer. This method may only be applied if all drivers at work can be reached through a despatching system. In the Montreal urban area, it was necessary to modify the method to include the many taxi customers using other ways to get a taxicab. The basic two-space stratification had to be nested with the different ways of obtaining passengers; phone calls, street cruising and taxi stand. The biggest problem was weighting the different strata. It was solved by doing special taxi counts on randomly selected street intersections and hourly counts of calls received by the taxi companies. Finally, the larger strata were weighted using a mathematical approximation. This method proved successful in determining first estimates of the supply and demand in the taxicab industry in the Montreal urban area.
Abstract