통근에서 승용차의 수단선택과 통행시간의 숨겨진 효과 : 개인, 가구, 거주지역의 다수준 결정요인을 중심으로

Abstract

Recent policy emphasis in the fields of urban and transport planning has been focused on reduction of vehicle miles traveled. Therefore, many empirical studies have tried to respectively identify either how mode choice private car is chosen against alternative transport ones or how much travel time by car is determined, not simultaneously both of them. However, such an unintentionally separated approach may give a bias on the estimation of actual private car travel time. This study is aimed at Identifying the unintentionally unrevealed impact of private car mode choice and its travel time in commuting, focusing on the multilevel determinant measures of individual, household, and residential area in the Seoul metropolitan region in Korea. Using the 2% sample raw data of the 2015 total population survey, the study employed the multilevel Tobit model to estimate private car commuting time with the latent variable of its mode choice against active modes. This is a combined model of the multilevel probit model for binomial mode choice and the multilevel linear regression model for private car commuting time. Comparing the two respective sub-models for mode choice and time of commuting, analysis results on the multilevel Tobit model demonstrate that there exist unintentionally but unrevealed effects in estimating actual private car commuting time in simultaneously considering individual behavior on mode choice and travel time. Such impacts are found in several determinant factors at all levels. The other conclusion coming from the analysis results is that reducing the probability of choosing private car on commuting is more effective than decreasing its commuting time from the perspective of actual consumed time.