Modeling car-sharing membership as a mobility tool: A multivariate Probit approach with latent variables

Henrik Becker , Allister Loder, Basil Schmid, Kay W. Axhausen, 2017, in Travel Behaviour and Society

doi:10.1016/j.tbs.2017.04.006
Location Switzerland
Population General
Sample size 8488
Factor analysis type exploratory factor analysis, varimax rotation
Stepwise regression nan
Removal of insignificant variables yes
Reviewed by NAH

Abstract

Individual travel behavior is to a large extent shaped by the respective portfolio of available mobility tools such as cars, season-tickets or a car-sharing membership. However, the choices of different mobility tools are interdependent and are also affected by individual attitudes. This paper presents an approach to jointly model the choice of four different mobility tools – including car-sharing. Using data from the Swiss transportation micro census of 2005 and 2010, it is shown that car-sharing is used as a supplement to a public transportation-oriented lifestyle, but is also used by car owners. The results further indicate that personal attitudes have a substantial effect on the choice of mobility tools and should therefore be accounted for when modeling such decisions. © 2017 Hong Kong Society for Transportation Studies

Factors

Variable Pattern loading
Road traffic revenues for PT infrastructure (Transport policy) 0.47
Road traffic rev. for slow mode infra. (Transport policy) 0.57
Road traffic rev. for environment (Transport policy) 0.56
Variable Pattern loading
Introduction of tunnel fees (Transport policy) 0.36
Introduction of peak hour fees (Transport policy) 0.55
Increase in parking fees (Transport policy) 0.64
Increase in fuel price (Transport policy) 0.63
Parking fees for shopping centers (Transport policy) 0.43
Road traffic rev. for road infrastructure (Transport policy) -0.3

Models

Dependent variable Mobility tool ownership
Model type Multivariate probit
Sample size 8477.0
R2 nan
Adjusted R2
Pseudo R2 (joint) 0.23
AIC nan
BIC nan
Log-likelihood at zero nan
Log-likelihood at constants nan
Log-likelihood at convergence nan
Car
Variable Coefficient p-value
age 0.005 <0.001
Male 0.372 <0.001
University degree (or equivalent) 0.196 <0.001
log household size 0.083 <0.01
log household income per sqrt household size 0.619 <0.001
log household income [kCHF] per householdsize 0.053 <0.001
Main accessibility 0.123 <0.001
Comp. better accessibility by transit 0.868 <0.001
Living and working in city center(s) 0.359 <0.001
Living and working an agglomeration 0.501 <0.05
Living and working in agglomeration or isolated city 0.39 <0.001
Living and/or working in a rural municipality 0.169 <0.001
Distance to closest car-sharing station [km] nan <0.01
Domestic distance travelled [100 km/day] 0.093 nan
Pro non-car infrastructure 0.145 <0.001
Pro traffic fees 0.304 <0.001
Indicator for 2010 census 0.039 nan
Constant 0.906 <0.001
Car-sharing
Variable Coefficient p-value
age 0.009 <0.001
Male 0.14 <0.05
University degree (or equivalent) 0.239 <0.01
log household size 0.016 nan
log household income per sqrt household size 0.036 nan
log household income [kCHF] per householdsize 0.056 nan
Main accessibility 0.079 nan
Comp. better accessibility by transit 0.195 nan
PT-microaccessibility at least level B 0.183 <0.05
Distance to closest car-sharing station [km] 0.007 nan
Pro non-car infrastructure 0.06 nan
Pro traffic fees 0.341 <0.001
Indicator for 2010 census 0.05 nan
Constant 2.378 <0.001
GA travelcard
Variable Coefficient p-value
age 0.043 <0.001
Agesq [y^2/100] 0.042 <0.001
Male 0.114 <0.05
University degree (or equivalent) 0.063 nan
log household size 0.052 nan
log household income per sqrt household size 0.011 nan
log household income [kCHF] per householdsize 0.128 <0.01
Main accessibility 0.415 nan
Comp. better accessibility by transit 0.252 <0.01
PT-microaccessibility at least level B 0.064 nan
Domestic distance travelled [100 km/day] 0.004 <0.001
Pro non-car infrastructure 0.111 <0.001
Pro traffic fees 0.446 <0.001
Indicator for 2010 census 0.009 nan
Constant 1.292 <0.001
Local Transit Pass
Variable Coefficient p-value
age 0.044 <0.001
Agesq [y^2/100] 0.04 <0.001
log household income per sqrt household size 0.013 nan
log household income [kCHF] per householdsize 0.104 <0.001
Main accessibility 0.034 nan
Comp. better accessibility by transit 1.207 <0.001
Comp. better accessibility to employment 0.48 <0.001
PT-microaccessibility at least level B 0.176 <0.001
Domestic distance travelled [100 km/day] 0.155 <0.001
Pro non-car infrastructure 0.241 <0.001
Pro traffic fees 0.337 <0.001
Indicator for 2010 census 0.139 <0.001
Constant 0.541 <0.001

The Attitudes and Travel Database is produced with support from the Center for Teaching Old Models New Tricks at Arizona State University, a University Transportation Center sponsored by the US Department of Transportation through Grant No. 69A3551747116.

sha256:a08d9e369743bf7e6d1c40d27347318209b40a7fb1543813fdcf31b898918815