Measurement of User's Preferences for Public Transportation through Computer Assisted Interviews

Citizens' preferences concerning bus transportation in Boulder, Colorado, were ascertained by two different but related techniques. The primary method was through a computerized game (’CARTFED,’ or Computer-Aided-Real-Time-Feedback Decision) in which subjects are asked to design a desirable bus system by choosing values for system characteristics. Subject's enter these in a portable computer-terminal and the computer calculates and reports the associated deficit, whereupon subjects reiterate making trade-offs to reduce the deficit to an acceptable level. In order to program the computer to calculate the deficits associated with, a wide variety of systems, it was necessary to estimate cost and revenue functions. Since the latter depends upon demand, a prerequisite to the CARTFED program-was the estimation of citizens' preferences by a different technique, which took the form of a questionnaire survey in which respondents were presented with randomly-generated profiles of possible bus systems and asked how often they would use such a system. Multiple regression analysis of the responses yielded the demand function necessary for the CARTFED program. This paper describes both techniques and their results.