How Do We Know Employer-Based Transportation Demand Management Works? The Need for Experimental Design

Abstract

Several studies of employer-based demand management programs have been conducted in the past few years. Many have drawn conclusions about effects of the programs by examining before and after program changes in the proportion of solo drivers among the employees involved in the program. Others compare before and after data to solo driving trends in the surrounding area or at selected employment sites. Still other studies do not use before and after data at all, but simply compare the proportion of solo driving among program participants after implementation with an average rate for the area or region. The weaknesses of these approaches to evaluating demand management programs are discussed, remedies for these weaknesses based on classical experimental design from the social science and statistical literature are suggested, and the results of an experimental design of employer trip reduction programs in the Denver region are highlighted. The example illustrates the required elements of an experimental design, as well as some difficulties in executing the design.