Economic and sociodemographic influences on autolessness: Are missing variables skewing results?
AD Gardenhire - Transportation research record, 1999 - journals.sagepub.com
AD Gardenhire
Transportation research record, 1999•journals.sagepub.comThe factors that influence whether a household will be without a vehicle, termed herein as
“autolessness,” are examined. First, national survey data in logistic regression models are
used to examine the strength of three variables identified in the literature as key influences
on autolessness: income, transit access, and urban residence. In a second model, however,
the inclusion of social and demographic variables representing race, sex, and age of the
household head and life-cycle status of the household reveals that they, too, exert …
“autolessness,” are examined. First, national survey data in logistic regression models are
used to examine the strength of three variables identified in the literature as key influences
on autolessness: income, transit access, and urban residence. In a second model, however,
the inclusion of social and demographic variables representing race, sex, and age of the
household head and life-cycle status of the household reveals that they, too, exert …
The factors that influence whether a household will be without a vehicle, termed herein as “autolessness,” are examined. First, national survey data in logistic regression models are used to examine the strength of three variables identified in the literature as key influences on autolessness: income, transit access, and urban residence. In a second model, however, the inclusion of social and demographic variables representing race, sex, and age of the household head and life-cycle status of the household reveals that they, too, exert substantial influence on the likelihood of a household being without a car. Furthermore, inclusion of these variables reduces the magnitude of the influence of the key variables. They exert influence along sociodemographic lines; for instance, female-headed households are more likely to be without a car than maleheaded households, controlling for the three key variables referred to above. However, these sociodemographic variables do not directly influence autolessness. It is hypothesized that they measure influences, such as savings, wealth, and access to credit, that are differentiated on sociodemographic lines. Inclusion of these variables in subsequent modeling efforts would explain many of the differences in autolessness attributed to sociodemographic variables in this model.