Charilaos Latinopoulos

A Tenuous Result: Re-analysis of the Link Between Internet Usage and Young Adults’ Driving License Holding. Comments on “Recent Changes in the Age Composition of Drivers in 15 Countries”

Sivak and Schoettle (2011) reported a cross-national comparison of trends in driving licencing rates among adults of different ages. The majority of the paper was devoted to presenting descriptive results. It was reported that in 8 of the 15 countries studied, there has been a decrease in driving licencing rates among young people (aged 20–24) and a corresponding increase among older adults, and in the other 7 countries there has been an increase in licencing rates among adults of all ages. Because crash rates vary strongly with driver age, there are potentially large implications for future trends in road casualties. The descriptive results are sound. The abstract states, however, that the primary implication of the study is based on a multivariate analysis: “The results of the analysis are consistent with the hypothesis that access to virtual contact reduces the need for actual contact among young people” (p. 126). This finding arises from the parameters estimated by linear regression with the various countries as the cases. Ten candidate country-level descriptors were used as independent variables. Stepwise linear regression with backward elimination, a data mining technique, was used to search among the 1,023 possible specifications. Four independent variables were retained in the preferred specification (gross national income at purchasing power parity, median age of the population, percentage of the population living in settlements of more than 1 million people, and number of Internet users per 100 people). Though the estimated regression parameters were not reported, it was reported that the number of Internet users per 100 (all age) population has a negative association with young adults’ rate of license holding with a t-statistic of −3.33. This result forms the basis for what is reported as the main implication of the study, namely, the negative association between Internet usage and young adults’ license holding.

Establishing the Links between Online Activity and Car Use: Evidence from a Combined Travel Diary and Online-Activity Pseudodiary Data Set

The links between online activity and physical mobility are of wide and growing interest to researchers, practitioners, and policy makers. This paper presents the results of an analysis of microdata from the Scottish household survey. The survey provides a unique, large-scale, nationally representative data set that includes both a travel diary instrument and a pseudodiary of participation in online activity. Multivariate regression models were estimated to relate peoples online-activity pro-files with their car driving mileage. The models included demographic and spatial characteristics to control for potential confounding effects. The analysis found that, net of other effects, compared with nonuse of the Internet, Internet usage was associated with a higher level of car use. The marginal effect of time spent online was, however, found to be negatively linked with car use. In other words, spending large amounts of time online is, all else equal, associated with less car driving mileage. The paper concludes with a discussion of further research needs to advance this line of inquiry.

Modeling Joint Charging and Parking Choices of Electric Vehicle Drivers: Decentralized Control Approach for Charging Service Providers

Electric vehicles (EVs) offer significant opportunities to improve sustainability of the road transport sector. But simultaneously, widespread adoption of EVs would create new challenges. For example, spatiotemporal concentration of charging events in high-density residential or commercial areas would place extreme demands on the power network, causing bottlenecks and grid instability. A novel approach to the typical decentralized control methods for EV charging service providers (CSPs) is presented. First, static price signals based on anticipated demand define a set of charging offers, targeted to segments of EV users. Prices are differentiated either only by time or both by time and place and allow comparison and evaluation of both scenarios. A choice-based revenue management method is employed to optimize allocation of generated charging offers, with respect to revenue outcome for the CSP. The charging coordination techniques are demonstrated through simulation. Data come from the London Travel Demand Survey and particularly trips around Westfield, one of Europes largest urban shopping malls, representing out-of-home charging behavior for short intervals in a high-demand area. Findings suggest that in a first-come, first-served system, locational pricing might create opportunities both for increased revenue and for relocation of charging events to less-congested facilities. In the revenue management system, locational pricing significantly favors total revenue outcome but without discharging vulnerable areas. However, because agents with conflicting interests participate in the process (infrastructure owners, power system operators, EV drivers), opportunity exists for the CSP to adapt constraints according to the priority of its objectives.