Walking is arguably one of the more dangerous modes of travel in the United States. A study by the Surface Transportation Policy Partnership (STPP) estimated that in 2001, the fatality rate per 100 million miles traveled was 20.1 for pedestrians, compared to only 1.3 for personal cars/trucks, and 0.8 for public transit. The vulnerability of pedestrians in motor vehicle-related crashes is higher in large central cities where large numbers of people walk in relatively compact urban environments. The National Safety Council estimates that 85.7 percent of all non-fatal pedestrian crashes and 72 percent of pedestrian fatalities in 2003 were occurred in urban areas.
The New York City Department of Transportation (NYCDOT) has identified pedestrian safety and mobility as high priorities and has devoted significant resources to pedestrian safety improvement programs. A variety of safety programs and initiatives have been implemented by NYCDOT over the past fifteen years. Such programs and initiatives appear to have been successful. According to a recent report by NYCDOT, overall pedestrian fatalities declined by 55.7 percent between 1990 and 2006. In 2005, of 32 cities in the United States with populations exceeding 500,000, New York was ranked the 10th safest city, with a pedestrian fatality rate of 1.85 per 100,000 persons. This number is lower than the average fatality rate (2.33) for those 32 cities, and significantly lower than cities like Jacksonville, Florida and Detroit, Michigan, which have fatality rates of 4.34 and 4.19, respectively.
Despite such improvements in pedestrian safety, pedestrians in New York City are still more vulnerable to motor vehicle-related crashes than those living in other parts of New York State or the United States. According to NYCDOT, the number of pedestrian fatalities as a percentage of total traffic fatalities has remained at roughly 50 percent since 1990. The share of pedestrian fatalities as a percent of total motor vehicle-related fatalities in 2005 was significantly higher for New York City than for New York State or for the United States as a whole. While this difference is not entirely surprising given that the rest of New York State and the nation also include non-urbanized areas, with less pedestrian exposure to motor vehicles, the uniqueness of New York City still stands out. Indeed, New York City is unique in many ways even when compared to other urban areas around the country. New York City is the most populated city, and the most densely populated city, in the United States. The population of the second largest city, Los Angeles (3.8 million), is less than half that of New York City (8.2 million). The population density of New York City (26,403 persons/mi2) is roughly 50 percent more than the next densest city (Patterson City, NJ – 17,675 persons/mi2); even San Francisco and East Los Angeles have population densities significantly below that of New York City (16,634 persons/mi2 and 16,698 persons/mi2, respectively). Moreover, there are more people walking in New York City than any other city. According to the 2005 American Community Survey’s journey-to-work data, 9.4 percent of New York City residents walked to work, while only 3.2 percent and 5.5 percent of the residents in Los Angeles and Chicago, respectively, walked to work. The result of these differences is a higher level of exposure for pedestrians in New York City than in other large central cities around the United States.
Project Overview
Task 1: Literature Review.
Task 2: Data Collection and Compilation. The key to the success of the proposed study is the development of an accurate and complete database of motor vehicle-related pedestrian deaths and serious injuries in New York City. In order to create this database, the team will cross-reference several datasets for the most recent five years (probably 2002 - 2006) that are made available or provided by NYCDOT and NYC DOHMH. Four types of datasets will be collected: fatalities, severe injuries, GIS layers (roadway characteristics, land use, transit network, etc), and socioeconomic data (e.g. census). As for fatality and injury data, Death Certificate/Vital Statistics, NYS MV 104A (source Police Accident Reports), NYPD Accident Investigation Squad Reports, and NYSDOT Safety Information Management System (SMS) files will be reviewed. In addition, the team will consider using other available datasets that potentially offer valuable information to support the data analyses. These datasets include the geo-coded Fatality Analysis Reporting System (FARS) provided to us by the National Center for Statistical Analysis NCSA, and the Multiple Cause of Death Files maintained by the Centers for Disease Control and Prevention. GIS layers such as street network, transit network and land use will be collected. Socioeconomic information is available from the Bureau of Census. Subtasks include development of criteria and categorization methodologies; development of geometric data from orthoimagery; geodatabase building; and data quality assurance.
Task 3: Descriptive Analysis. To develop a correlation between accident type and intersection type/corridor type, the team will conduct a comprehensive analysis of factors contributing to pedestrian crashes. Using the datasets compiled in Task 2, this task will provide comprehensive analyses of the fatal and serious injury pedestrian crashes. As part of these analyses we will describe the frequency and location of pedestrian crashes; determine crash trends; and establish correlations between accidents and intersection/corridor types. The results of the analyses will be summarized using tables, graphs, matrices and maps. The data will be broken out by the following attributes: crash locations; demographics of the victim and driver; street type where fatality occurred; time of crash; roadway conditions; types of injuries; death by type of vehicle; pedestrian deaths and serious injuries as percent of total accidents; motor vehicle operator characteristics and contributing factors; roadway characteristics; land use; and geographic/administrative boundaries. The team will then geocode the crash locations and use spatial aspects of the data to conduct a more in-depth geographic analysis. The spatial analysis will include statistical examination of clustering patterns.
Task 4: Crash Cause Modeling. This task will analyze the relationship between pedestrian crashes and other independent variables such as roadway-geometry (transportation environment variables) and socio-demographic variables of the pedestrians/drivers. This will be performed using the comprehensive database compiled in Task 2 and the analysis using GIS maps conducted in Task 3. Different severity models will be developed for pedestrian crashes which are classified into fatal and injury crashes. The analysis will consider variables associated with vehicle characteristics, traffic and roadway-geometric features, driver characteristics, pedestrian characteristics and environmental characteristics. The identification of additional variables into the modeling process will be updated based on the input from the NYCDOT staff, and coordinated with the existing typology of NYCDOT. Using alternative modeling techniques (such as binary logistic regression, ordinal logit and multinomial models) this study will identify relationship between crash severity and the crash-related variables based on the intersection and corridor type. This analysis will identify controllable factors based on intersection and corridor types that mainly contribute to the occurrence of pedestrian crashes in New York City. The pedestrian crash risk analysis will include two distinct components: 1) crash frequency analysis; and, 2) severity or crash consequence analysis.
Task 5: Findings and Recommendations.