Walkability as a Summary Measure in a Spatially Autoregressive Mode Choice Model: An Instrumental Variable Approach

In recent years it has become more common to include social interactions or neighborhood effects (also called social network effects) in transportation modeling. These models are typically in the tradition of Brock and Durlauf (2001, 2002) who were among the first to propose a discrete choice model that includes social interactions and neighborhood effects. However, their approach is inherently non-spatial, while the topology of social interactions and neighborhood effects can be best captured spatially (Leenders 2002; Paez et al. 2008a). Therefore, some of the latest articles in transportation modeling have moved towards explicitly incorporating the spatially autoregressive structure of social network effects into their models (e.g. Dugundi and Walker 2005; Paez and Scott 2007; Goetzke 2008). This new direction in transportation research is not all that surprising, given the success of spatial econometrics as an emerging modeling method across social science disciplines. The econometric strategy to implement an independent variable representing social interactions and neighborhood effects, as proposed by Brock and Durlauf (2001, 2002), is to use the group mean of the observed dependent choice variable, as defined by social interactions and neighborhood effects. This approach can be spatially extended if the group mean is based on spatial relations, as in traffic analysis zones (Dugundi and Walker 2005), a spatial weight matrix (Goetzke 2008), or a matrix based on personal relations (Paez and Scott 2007). Therefore, choices could be modeled as a function of the typical choice determinants in travel behavior analysis (e.g. personal, household, trip and mode characteristics), or as choices of either a non-spatial group or spatial neighbors. Empirical work dealing with mode choice decisions by Dugundi and Walker (2005), and Goetzke (2008) give evidence that the mode choice decisions of spatial neighbors are indeed associated with the mode choice decision of the individual.

[1]  William A. Brock,et al.  Discrete Choice with Social Interactions , 2001 .

[2]  K. Train.  Discrete Choice Methods with Simulation , 2003 .

[3]  Mark M. Fleming.  Techniques for Estimating Spatially Dependent Discrete Choice Models , 2004 .

[4]  Kenneth E. Train,et al.  Discrete Choice Methods with Simulation , 2016 .

[5]  W. Evans,et al.  Measuring Peer Group Effects: A Study of Teenage Behavior , 1992, Journal of Political Economy.

[6]  Roger Th. A. J. Leenders,et al.  Modeling social influence through network autocorrelation: constructing the weight matrix , 2002, Soc. Networks.

[7]  Joan L. Walker,et al.  Discrete choice with social and spatial network interdependencies: An empirical example using mixed generalized extreme value models with field and panel effects , 2005 .

[8]  L. Anselin.  Spatial Econometrics: Methods and Models , 1988 .

[9]  Frank Goetzke.  Are Travel Demand Forecasting Models Biased because of Uncorrected Spatial Autocorrelation , 2003 .

[10]  A. Páez,et al.  A Discrete-Choice Approach to Modeling Social Influence on Individual Decision Making , 2008 .

[11]  Darren M. Scott,et al.  Weight matrices for social influence analysis: An investigation of measurement errors and their effect on model identification and estimation quality , 2008, Soc. Networks.

[12]  Frank Goetzke,et al.  Network Effects in Public Transit Use: Evidence from a Spatially Autoregressive Mode Choice Model for New York , 2008 .

[13]  J. Brueckner,et al.  Social Interaction and Urban Sprawl , 2006, SSRN Electronic Journal.

[14]  W. Brock,et al.  A Multinomial-Choice Model of Neighborhood Effects , 2002 .

[15]  P. Zarembka.  Frontiers in econometrics , 1973 .

[16]  Jeffrey M. Wooldridge,et al.  Introductory Econometrics: A Modern Approach , 1999 .

[17]  Sergio J. Rey,et al.  Advances in Spatial Econometrics: Methodology, Tools and Applications , 2004 .

[18]  A. Páez,et al.  Social Influence on Travel Behavior: A Simulation Example of the Decision to Telecommute , 2007 .

[19]  Kenneth C. Land,et al.  On the Large-Sample Estimation of Regression Models with Spatial- Or Network-Effects Terms: A Two-Stage Least Squares Approach , 1992 .

[20]  D. McFadden.  Conditional logit analysis of qualitative choice behavior , 1972 .