Using Orthogonal Arrays to Design State-Choice Surveys in Transportation Planning

摘要

This article outlines a procedure to develop a factorial stated-choice survey that recognizes that goods and services have value because of their attributes (product characteristics). A systematic and planned design process is followed after predefining the attributes and varying their levels without measurement error to create choice (preference) alternatives. As an illustration, the procedure described in this article will be applied to transportation planning. A questionnaire is developed to get information from current and potential users of a HOV (high occupancy vehicle) lane with excess capacity being transformed into a HOT (high occupancy toll) lane. Conceptually, the process followed in the design of the stated-choice survey includes these tasks: 1. Selection of the attributes that affect a product or service, and appropriate levels (values or options) for each one. 2. Identification of relevant choices. 3. Transformation of attributes and options into factors. 4. Design a fractional factorial to generate the questions included in the survey. 5. Selection of a blocking strategy to divide the entire questionnaire into groups of questions that can be answered by an individual respondent. 6. Generation of a catalog of orthogonal arrangements from which to select a most appropriate design for each particular application.