AN ANALYSIS OF HOUSEHOLD VEHICLE TYPE ACUQISITION USING MULTINOMAIL LOGIT MODEL

Abstract

The US is a highly motorized society and personal vehicles are clearly an integral part of American life in that people are used to rely on their personal vehicles when both traveling and commuting. According to the 2001 National Household Transportation Survey (NHTS), household travel generates more than 80 percent of the total vehicle miles of travel in the United States and 87% of the daily trips are made by personal vehicles. Overall personal vehicles have been outnumbering licensed drivers since 1972 at an ever increasing rate. Each year more than 100 new mainstream models are manufactured and millions of vehicles are sold in America. Naturally, people ponder which type of vehicle to select based on their own preferences and needs out of a wide range of makes and models. It can be varied with household demographics, residence location, or primary driver?s characteristics. Historically, there has been a trend of a popular type of vehicle in various time spans such as compact cars in the mid-1970s, minivans in the 1980s, and pickups/sport utility vehicles in the 1990s. Here, an important question arises; what factors determined or contributed to the choice of a certain type of vehicle? In order to find this kind of question, many researchers have devoted to identifying the factors that have an influence on peoples? vehicle choosing behaviors. As a result, a variety of models have been developed for explaining the vehicle type choice and disaggregate choice models such as multinomial logit model (e.g. Lave and Train, 1977; Mannering and Winston, 1985; Choo and Mokhtarian, 2004) and nested logit model (Berkovec, 1985; Feng et al., 2005; Mohammedian and Miller, 2003) for the vehicle type measure and continuous regression model for the vehicle use measure have been widely used by many eminent studies in this field. These models generally focused on several types of determinants such as vehicle attributes, household 2 demographics, residence location factors, primary drivers? characteristics, or personalities and travel attitudes. Several earlier models with such those general factors basically show consistent results, however, failed to capture the unique features except for the intuitive results whatever methods they used. Of course, it is true that the earlier studies produced undoubtedly remarkable achievements with providing important insights into the factors affecting the choice of vehicle type from their perspectives.