THREE ESSAYS ON URBAN TRANSPORTATION STUDIES IN WASHINGTON D.C.: SAFETY EFFECT OF ALL-WAY STOP CONTROL, SAFETY EFFECT OF REVERSIBLE LANE AND LOADING ZONE ALLOCATION
MetadataShow full item record
Systematic data-driven and evidence-based urban transportation policy making and evaluation become increasingly important for public agencies to ensure transparent and efficient services. This dissertation, consisting of three essays on urban transportation studies, focuses on two issues (safety and asset management) that are broadly related with urban transportation policy making and evaluation in Washington D.C. In Chapter One, I evaluate the safety effect of All Way Stop Control (AWSC) conversion with an observational treatment group and a randomly selected control group from stratified samples. Selection bias and time trend are controlled using empirical strategies such as Multiway ANOVA and Difference-in-Differences analysis. The study reveals statistically significant reductions of right angle crashes upon AWSC conversions. However, for all the other collision types, including right turn, left turn, rear end, sideswipes and bicycle crashes, none of the estimated coefficients were statistically significant. In addition, the study quantified a statistically significant increase of straight hit pedestrian crashes upon AWSC conversion. In Chapter Two, I study the safety effect of removing reversible lane operations along urban arterials. Taking advantage of the termination of three reversible lane arterials in 2010, the evaluation is performed using the Before-After (BA) study with a control group and the Empirical Bayes (EB) method, respectively. I estimate Crash Modification Factors (CMF) for all crashes, fatal/injury crashes, property damage only (PDO) crashes, rear-end crashes, left turn crashes and sideswipe crashes. My findings suggest a clear tradeoff between safety and the gain of peak direction capacity by operating reversible lanes along urban arterials. In Chapter Three, I propose an innovative procedure for allocating scarce curbside space for loading zones in an equitable, quantifiable and repeatable manner. Freight Trip Generation (FTG) models are used to estimate the delivery needs for business establishments at a block face level. The current numbers of loading zones per block face are regressed against the Gross FTG (GFTG) per block face and other block face characteristic variables using zero-truncated Negative Binomial models to establish a baseline. Curbside spaces are then assigned as loading zones in an iterative process.