Center Identification Number: 416-02
Project Title: Pedestrian Mid-block Crossing Difficulty
Xuehao Chu, Senior Research Associate
Michael R. Baltes, Senior Research Associate
Center for Urban Transportation Research
External Project Contact:
I. Project Objective
To develop a level of service measure for mid-block crossing of roads by pedestrians. This measure is intended to indicate pedestrians' perceived level of difficulty in crossing roads. In the hierarchy of FDOT's family of multi-modal level of service measures, ranging from areawide, corridor-based, facility-based, segment-based, and point-based, the one proposed in this project is a point-based measure. The study will attempt to determine what variables are correlated with such pedestrians' perceived crossing difficulty. This will be done through a statistical calibration and validation process, involving collecting actual site characteristics and stated levels of crossing difficulty by a sample of persons at a sample of sites. These variables will include those that are most important to the Department and local governments for the purpose of improving mobility, safety, and livability.
II. Project Abstract
The Florida legislature passed HB-17, the Urban Infill and Redevelopment Act amending F.S. 163.3180 to state that "Local governments should be specifically encouraged to employ alternative techniques for measuring level of service, including multi-modal, vehicle miles traveled (VMT)-based, access-based, and zone-based approaches." It directs further that: "FDCA and FDOT should work aggressively to provide technical assistance to local governments to employ these preferred level of service methodologies."
Currently there are no known techniques for quantifying the level of service for pedestrians crossing streets or intersections. FDOT's present multi-modal level of service research efforts measure facility segments or a designated area. In determining bus level of service a street crossing difficulty factor has been incorporated. This measure has not been tested on a real world population. It is derived through automobile level of service data such as AADT, number of lanes, presence of a restrictive median, speed and the distance between signalized intersections. This indicator will not describe the crossing difficulty perceived by pedestrians at specific locations. With more than 47.8 percent of Florida's pedestrian fatalities occurring while attempting to cross a street at mid-block, it is evident that a reliable measure is needed which can describe the pedestrian crossing environment.
III. Task Descriptions
The project is expected to be carried out in two phases. The first phase will focus on research design, including the identification of site characteristics, sampling design, data collection procedure, statistical analysis procedure, and the development of an abstract spreadsheet model for future application. The second phase will focus on carrying out the research, including data collection, statistical calibration and validation, parameterizing the spreadsheet model, and preparing the final report for the entire project. The two-phase structure coincides with the two-year funding phase of the NCTR.
PHASE ONE - YEAR ONE
Task 1 Review Literature
The research team will review the literature related to pedestrian crossing difficulty, including the search developed by the University of Florida for FDOT and other sources. In addition, the team will collect documents provided by FDOT and its researchers on their current efforts in developing multi-modal level of service measures. The research team will get a full understanding of the Department's program on multi-modal level of service measures, the role of the pedestrian crossing difficulty measure in the program, and how the various components in the program fit together. The team will also review the literature on site characteristics that are important determinants of pedestrians' perceived crossing difficulty.
Task 2 Assemble Advisory Committee
The research team will assemble an advisory committee for the project. It is envisioned that the committee will involve people from the Department (including Systems Planning and Bicycle/Pedestrian Program), the Department's consultants on multi-modal level of service measures, and people from local agencies in the study area (including MPO, municipal government). The main function of the committee will be to provide inputs to the research team on its research design. The committee may also participate in the monthly teleconference calls. The committee and the researchers will have a kickoff meeting in Tampa once it is formed.
Task 3 Identify Site and Volunteer Characteristics
The research team will identify site characteristics in several steps. The team will start with a preliminary set, including distance between the nearest signalized intersections, crosswalks, bicycle lanes, bus shelters, lighting, signs, number of lanes, volume at time of surveying, median treatments, and landscaping. The second step will refine this preliminary set with information collected from the literature review. For example, a 1993 article by Hunt and Abduljabbar in Traffic Engineering and Control 34 (11), "Crossing the Road: a Method of Assessing Pedestrian Crossing Difficulty," suggests that traffic platooning is critical in determining pedestrian crossing difficulty. The third step will identify those factors that are expected to affect pedestrian crossing difficulty and are most important to the Department and local governments for the purpose of improving mobility, safety and livability. If necessary due to lack of information from the literature, for example, the research team will also conduct focus groups to help identify site characteristics. For volunteer characteristics, the research team will rely on the literature and experience on what personal and household characteristics are likely to influence pedestrian crossing difficulty.
Task 4 Determine Sampling Approaches
The research team will determine the sampling approaches for both sites and volunteers. This will include the sampling frames, sample sizes, and specific sampling methods. On sampling frames, the team expects to select one metropolitan area (e.g., Tampa Bay). Sample sites will represent a full range of roadway characteristics. High pedestrian crash locations shall be included as test sites. Volunteers will be limited to local residents and will represent a full range of personal and household characteristics. Sample sizes will depend on the number of site characteristics to be included in the study and how many sites each volunteer will cover. The sampling methods will need to specify how exactly each volunteer or site will be selected from the corresponding sampling frame.
Task 5 Specify Data-Collection Procedure
This task will specify the procedure to be used to collect data once site and volunteer characteristics are identified and sampling approaches are determined. For each of the site characteristics, the team will describe the data requirements, data source, and methodology for calculation. For volunteers, the team will design two survey instruments. One of them will be used to collect personal and household characteristics of the volunteers, including whether they are frequent walkers. These characteristics can be important determinants of how people perceive crossing difficulty and are critical to isolate them from site characteristics. The other instrument will be a stated-preference survey. One purpose of the stated-preference survey is to solicit the volunteers' perception of the crossing difficulty at specific sites, given the observed environment. In addition, the survey will specify one of several common reasons for crossing a road and will solicit whether the volunteers would cross the road under this hypothetical case. These reasons could include such activities as catching a bus, living at the other side, going to a favorite restaurant, etc. In addition, the team will also specify how exactly the surveys will be conducted. This question of hypothetical crossing purposes will be conducted so as to not bias questions on the perceived crossing difficulty. As part of this task, the research team will conduct an informal pilot test of these surveys on a small number of people recruited at the USF campus.
Task 6 Develop Statistical Procedure
For future application, each of the site characteristics need to have a weight attached that reflects its relative importance in influencing crossing difficulty. To accomplish this, one needs a mathematical framework to connect these site characteristics to crossing difficulty and a statistical approach to determine the relative weights within that mathematical framework. The choice of the mathematical framework should have a behavioral foundation. The statistical approach allows the weights for the site characteristics to be estimated. Such a statistical approach should be chosen with care so that the weights are estimated with precision and accuracy.
Task 7 Develop a Spreadsheet Model
Based on the above tasks, the research team will specify an expected model that would link site characteristics to a meaningful measure of perceived crossing difficulty. The team will then implement this expected model into a spreadsheet template. This will take into account parallel spreadsheet templates for the other components of the Department's multi-modal level of service measures. The team will document the model process, required input data, and appropriate uses of the model. The model will be provided to the Department in an electronic file. It is expected that the draft spreadsheet will be delivered in early September 2000. The research team will modify the model if necessary upon receiving the comments from the Department.
Task 8 Develop Issue Paper
The research team will document the research design reflected in the above tasks into a draft issue paper. It will outline the various approaches that the Department could use to establish a pedestrian crossing difficulty measure, including the pros and cons of each approach. This draft paper will be sent to the FDOT project manager and members of the Advisory Committee for input. The research team will then modify the research plan and the issue paper, taking into account these inputs. This finalized issue paper will serve as the blue print for Phase Two and the final report for Phase One. An electronic file and ten hard copies of the finalized issue paper will be sent to the FDOT project manager.
PHASE TWO - YEAR TWO
Task 9 Select Sites
The research team will select sites according to the research design documented in the issue paper. It is important that the selected sites are representative of the sampling frame and have enough variation on each of the site characteristics. One way to accomplish this would be to assemble a computerized data base on roadway segments with information on key characteristics such as the number of lanes and AADT and then to sample segments within each stratum defined by different combinations of these key characteristics.
Task 10 Collect Site Characteristics
The research team will collect data on the pre-determined set of site characteristics from the selected sites according to the research design documented in the issue paper from Phase I. It is important to separate static from dynamic characteristics. Static ones do not change temporally over a short period, while dynamic ones do. The number of lanes would be an example of a static characteristic. Traffic flow and traffic platooning would be examples of dynamic characteristics. Static ones can be collected before the stated-preference survey on site. Dynamic ones should be collected at the time of the stated-preference survey on site.
Task 11 Recruit Volunteers
The research team will recruit volunteers according to the research design documented in the issue paper from Phase I. It is likely that the volunteers will be limited to persons age 13 years or older. Ideally the volunteers should be selected from the general population through a sampling process so that they are representative of the population. It could be logistically difficult to do so. An alternative would be to recruit sponsors first and then select their employees or members as volunteers. The downside of this alternative is that the sampling frame would be limited to people within these sponsors, which are not likely to be representative of the general population.
Task 12 Conduct Volunteer Surveys
The research team will conduct the two surveys on the recruited volunteers according to the research design documented in the issue paper from Phase I. This involve at least two steps. Step one will gather all volunteers to a central location for a quick training session. Each volunteer will then complete the survey on personal and household characteristics. Also, a guide is assigned to a group of several volunteers. Step two will carry out the stated-preference survey with each of the volunteer groups at pre-selected sites for them after leaving the central location. Volunteers will be accompanied by the assigned guide to a selected site and will rate this site as to the difficulty in crossing the roadway and will indicate their crossing decision. They will not actually step into the roadway. They will complete the survey and give it to the guide. They will then proceed to other pre-determined sites and repeat the procedure.
Task 13 Statistical Calibration and Validation
The research team will use a large portion of the data set to calibrate the expected model, according to the pre-selected mathematical framework and statistical approach. Alternative specifications of the model, including variations in the variables, the mathematical framework, and the statistical approaches, will be tested. The team will then use the other portion of the data set to test the accuracy of the alternatives. That is, the values of the site and volunteer characteristics will be used in the model to calculate the predicted crossing difficulty and then compare the predicted ones with the stated ones. This will be done for alternatively estimated forms of the model. These forms can then be compared in terms of their prediction errors.
Task 14 Prepare Final Deliverables
The research team will specify the spreadsheet model with parameters estimated from Task 13 and will prepare the final report. The final report will include not only the research design contained in the issue paper but also data collected, lessons learned, and results from the statistical calibration and validation from Phase II. The final report will also include the research team's recommendations on further improvements and testing on the pedestrian crossing measure. An example of further testing would be to examine the transferability of the model developed with data from one metropolitan area to other areas. It is expected that draft final deliverables will be available by the end of June.
IV. Student Involvement
Students will be used to compile data and assist in the review of progress reports and summary of projects' strengths and weaknesses.
V. Relationship to Other Research Projects
This project is part of FDOT's research efforts on multi-modal level of service measures. Related projects include those in the University of Florida multi-modal contract, the University of Florida area-wide analysis in Gainesville, and RS&H area-wide analysis in Orlando, as well as FDOT multi-modal level of service projects in other districts.
VI. Technology Transfer Activities/Peer Review
The results of this analysis will be provided to the FDOT through a series of quarterly progress reports, an issue paper on methodology, a spreadsheet model, and a final report. Copies of the model and the final report will be widely distributed to various agencies, offices, and decision-makers in Florida.
VII. Potential Benefits of the Project
The results of this research will provide transportation and land use professionals with a tool that can assess the degree of difficulty a pedestrian will experience crossing a roadway. Typically the fatal flaw in our non-automobile transportation system occurs when it is impossible for the user to cross roadway. The research results will allow a what-if analysis to determine the effect of proposed improvements. This research will be used by local governments, MPOs and FDOT in:
This research will provide a tool to help implement the Urban Infill and Redevelopment Act of 1999. As Florida is a national leader in multi-modal level of service, this research will serve as a national model.
National Center for Transit Research · at the Center For Urban Transportation Research · University of South Florida · 4202 E. Fowler Ave., CUT100 · Tampa, FL 33620-5375 · (813) 974-3120 · (813) 974-5168 · www.nctr.usf.edu · Comments: email@example.com