Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
8 The introduction makes the case for a performance-based management approach for pedes- trian safety and describes the overall steps needed to obtain and utilize the necessary types of data, and to develop the tools and processes to achieve a performance-based practice similar to what is commonly used in road safety management programs for other modes of travel. The first step of the systemic process is to define the study scope. This step sets the stage for all subsequent steps. This step outlines basic procedures to identify a focus (i.e., a âtargetedâ risk problem) for a systemic pedestrian safety analysis, regardless of the agency or network size. These include â¢ Defining the jurisdiction or network area for analysis; â¢ Identifying one or more target facility or location types; and â¢ Identifying subsets of target crash types for systemic focus. C H A P T E R 2 Step 1: Define Study Scope Definition: Network In this guidebook, network refers to the complete network of streets within a defined area or jurisdiction. See the considerations in the following section that may help to determine the network to include in an application of a systemic process. Identify Network for Analysis The jurisdictional focus may be self-evident, depending on who is initiating the process and how later steps in the process may be shared or divided. For example, a state DOT may initiate a high-level assessment of pedestrian crash issues across the state or for different geo-political divisions. These analyses can, in turn, be used to inform decisions about regional priorities for further systemic safety analyses. At a statewide or regional level, the network of interest (and high-level analysis) might first be subdivided or stratified by ownership or area-type. Examples include the following: â¢ Urban (municipal) versus rural locations â¢ State, county, city, and other network divisions Urban and rural location and facility types can be analyzed systemically but there may be issues regarding compiling certain types of supplementary data. Compiling crash data from
Step 1: Define Study Scope 9 different regions within a state should also be feasible, but there may be more challenges during Step 2 when compiling land use, census, or transit dataâtypes of data that are highly desirable for an analysis of pedestrian risk factors. Also consider differences in regional travel patterns, geographies, development types, and other characteristics when deciding on a net- work focus. A state or regional agency may use area-based analyses to help identify higher-risk focus areas for more in-depth systemic analysis. Performance measures such as crash frequencies, crash rates per population, and crash rates derived from travel estimation surveys that capture pedestrian trips or commute mode-shares may be summed and used to compare risk at regional scales. See FHWAâs pending Guide for Scalable Risk Assessment Methods for Pedestrians and Bicyclists (Turner et al., in press) for more information on performance measures and their use in this type of analysis. Once the network focus has been determined, the next tasks focus on identifying a target facility and crash type that can be used to assess risks at specific locations for potential treat- ment. If there is a strong trend of crash risk at an area scale, such as by distinct neighbor hood characteristics, various land uses, or population characteristics (such as older or school- aged pedestrians, any type of impairment, or others) and there is an intention of conduct- ing more field-based diagnoses to identify risk patterns for systemic treatment within zonal areas, an area-based scale could be used for analysis. In general, however, most agencies will prefer a target roadway location type such as intersections or segments, since this allows for identification of risk characteristics associated with specific locations that may warrant treatment. Looking Ahead It is a good idea to identify an analysis network that has clear boundaries within which an agency and partners have the authority to make changes, since the goal is to implement treatments. The data types needed for risk analysis are also more likely to be available and readily compiled for cohesive urban or rural regions or municipalities that tend to compile land use and other types of planning data. Identify One or More Target Locations and Crash Types The purpose of identifying a specific target crash type or types is to narrow traffic safety efforts down to target crash types that lend themselves to more readily identifying risk patterns and potential treatments. Prioritizing a specific crash type in the context of improving pedestrian safety enables engineers and planners to identify risk factors that influence those specific types of crashes, which can lead to better targeted, systemic treatments. The target facility and crash type can be informed by prior studies or determined by analyses. Determine if there are pre-existing safety planning documents that have defined key pedestrian focus issues. These may include strategic highway safety plans or, at a more local level, summary analyses of pedestrian crash problems that may have been developed for pedestrian safety action plans or other planning documents. Otherwise, turn to crash data to help define the target facility and crash types.
10 Systemic Pedestrian Safety Analysis Key Crash Data Elements Needed to Identify Crash Targets Crash data elements that are key to a systemic pedestrian safety approach include the following: â¢ Crash type: Crash types describe the events leading up to a crash and summarize the conflict type or relative approach angles, positions, and maneuvers that led up to a collision. Crash types for motor vehicleâonly collisions tend to be well defined and these provide information relevant to treatment decisions. Many resources treat pedestrian crashes as if they are all the same but, in fact, different circumstances and maneuvers are present at different âtypesâ of pedestrian crashes, just as for motor vehicleâonly crash types. This information is useful in a systemic pedestrian safety process, as it begins the work of diagnosing patterns systemically that are widespread and potentially treatable. â¢ Location type: The location type is either at an intersection or segment location or potentially at another location type. Where the crash occurred is also relevant for a systemic analysis process since this information helps to determine crash context and treatment possibilities. Location information should be available in crash data. Compared with arterial classifica- tions or other potential roadway descriptors that might be used in a motor vehicleâfocused approach, focus on location characteristics most relevant for identifying risks and potential systemic pedestrian safety measures. â¢ Injury or crash severity: A systemic approach aims to target treatments to locations with higher potential for severe injury crashes. Thus, pedestrian injury or crash-level injury indicators are also potentially useful. However, as discussed below, any pedestrian crash is potentially severe. Definitions: Crash Type and Crash Frequency A crash type is a variable that typically describes events and maneuvers of the involved parties that led up to a crash. The relative maneuvers of the parties such as road departure (single vehicle), angle crash (between two motor vehicles), or pedestrian crossing at midblock and struck by a vehicle traveling straight (pedestrian-motor vehicle crash type) are examples. Crash frequency is the number of crashes at a defined location over a defined time. Noteworthy Practice Arizona coded all pedestrian crash types for crashes that occurred on the state highway system for their pedestrian safety action plan, which included both high crash and systemic/risk-based assessments. See Case Example 3 for details. In Seattle, pedestrian crash types were unavailable. However, there were separate variables available on the motoristsâ and pedestriansâ pre-crash maneuvers. A cross-tabulation of these two variables was used to generate a pedestrian crash type. This ââtypeâ was used to determine high frequency scenarios for systemic focus. See Case Example 1 for more information.
Step 1: Define Study Scope 11 Crash Analysis to Identify Focus Types Begin the analysis with a careful inspection of crash frequencies and tabulations to identify high frequency location and crash type combinations. The goal is to narrow down to one or several target crash types for more in-depth risk analysis in Step 3. At a minimum, state or regional agencies may wish to develop crash frequency tables and cross-tabulations or a crash tree to subdivide or stratify the crash data on factors such as those previously mentioned to summarize the frequency and proportions of crashes by the following: â¢ Crash location characteristics â Rural or urban (if relevant) â Intersection or non-intersection (segment) â Signalized or unsignalized crossing location â¢ Crash type (e.g., those involving pedestrians) Cautions and Considerations Other elements can be used to further subdivide the data, such as type of traffic control or other roadway characteristics. Use caution, however, when creating subsets with a narrow focus. If the focus location or crash type is too narrow, the sample of pedestrian crashes and locations may be too small to meaningfully identify risk factors. Examples of Target Crash Types Most prior studies have focused on certain types of locations such as intersections (sometimes only signalized intersections) or segments and have been less likely to also subdivide by crash Troubleshooting If crash type information is unavailable, here are some options to consider. â¢ Consider coding crash types for all pedestrian crashes. This data gap can sometimes be solved by using a standard crash-typing system for pedestrian crashes, such as the Pedestrian and Bicycle Crash Analysis Tool (PBCAT). â¢ Explore whether a combination of variables that are available in the crash databaseâsuch as motorist and pedestrian positions, maneuvers, or actionsâcan be used to characterize key pedestrian crash types. â¢ If neither of these options is workable, then there may be a need to consider whether improvements can be made in crash reporting for future efforts. Most states collect several separate crash types for motor vehicleâonly collisions, and this information is useful for systemic safety programs. The goal could be to identify a limited number of pedestrianâmotor vehicle crash types that are relevant for identifying conflict patterns and add these to crash reporting. Remember that crash types, like motorist types, describe pre-crash maneuvers and events, and are in addition to any behavioral contributing factors (speeding, failure to yield, or impairment) on the part of either the driver or pedestrian.
12 Systemic Pedestrian Safety Analysis types. However, crash type data may be useful to identify risk factors and locations that may be most in need of treatment. The prior tendency to use location types, and not crash types, may be due to the frequent lack of crash type information in pedestrian crash data. Consider focusing on a few basic crash type descriptors such as the following: â¢ Motorist traveling straight strikes crossing pedestrian â¢ Motorist turning left strikes crossing pedestrian â¢ Motorist turning right strikes crossing pedestrian â¢ Nighttime crashes â¢ Fatal and injury crashes Other crash type descriptors besides those listed above could also be used. For example, child pedestrian crashes could potentially be identified as a target crash type. When selecting the sys- temic focus types, keep in mind whether they relate to application of potential systemic counter- measures and whether the issue is well defined in pedestrian crash data. In terms of pedestrian crash severity, it is recommended that all pedestrian crashes be used or all fatal and injury crashes. This allows an adequate sample of crashes for analysis. Later, more options are provided on ways to consider risk factors in the analysis that increase potential for severe injuries, as well as when comparing treatment options. Looking Ahead In Step 2, roadway characteristics from roadway inventory files are compiled for analysis. However, if crash and roadway data are already linked, the location characteristics from roadway inventory can be used in this step as well. As crash data are being reviewed, this is also an opportunity to determine the availability, quality, and completeness of specific crash location data. In other words, determine whether detailed location descriptions such as latitude or longitude coordinates, or other spatial linking variables, are available to identify the specific roadway locations for where each crash occurred. If latitude or longitude or unique location identifiers are attached to the crash data, bring these variables into the crash database with the location, crash type, and other descriptors you are using to identify the target type. These variables will be used later to count the crashes by location for a network-based risk analysis. Identifying a target crash type also suggests thinking about the list of potential effective countermeasures, where these might be applied, and the crash types they treat. Step 5 will provide more information about potential countermeasures for a systemic pedestrian safety process. At a statewide level, the determination of the systemic pedestrian safety focus may begin with categorizing whether crashes occurred within municipalities or outside municipalities and drill down to basic location types. See Figure 4 for an example of a crash tree diagram to identify potential target area, location, and roadway types using data from North Carolina. In this exam- ple, non-intersection crashes on two-way, undivided roads are candidate target location types
Step 1: Define Study Scope 13 for further systemic analysis for both rural and urban locations statewide. Two-way, divided roads also account for sizable numbers in urban and rural locations. Other metrics that normalize crash frequencies with an exposure denominator (estimates of pedestrian trips, mode share, or other) can also be used to help select a network or networks for further systemic analysis. Within a city or urban region, the focus could begin by assessing the frequencies or propor- tions of crashes by location type, then the most prevalent crash types at each location. Consider- ation can be given to total frequencies and proportions of severe crashes accounted for by each subset. See Case Example 1 for an illustration. Noteworthy Practice North Carolina Department of Transportation (North Carolina DOT) sponsored multi-year projects to code pedestrian and bicycle crash types, geo-locate each pedestrian and bicycle crash that is reported statewide, and add these elements to existing crash variables. The crash data are available in a queryable database (i.e., the North Carolina Pedestrian and Bicycle Crash Data Tool available at http://www.pedbikeinfo.org/ pbcat_nc/), so initial systemic analyses can easily be performed to identify crash type and other crash patterns. Crash maps that include crash types and many other crash characteristics are also available for exploration, and data are available for local agencies to spatially link and use for safety analysis. Figure 4. Example tree diagram of pedestrian crash location types using North Carolina crash data.
14 Systemic Pedestrian Safety Analysis Finalize Area and Location Type Scope There may be a need to reassess the network focus extent, road types, or potentially crash types for the risk analysis after compiling the other data types for the risk analysis in Step 2. For example, there may not be traffic volume data for the whole network within the desired focus area. There may be opportunities to collect or enhance the needed data types or refine the focus for the initial analysis. The process will provide additional lessons and knowledge that will be useful for the next round and for improving and sustaining the systemic approach to pedes- trian safety. Oregon DOT initially planned to focus on both state and local roads, but due to a lack of consistent roadway inventory data for local roads, the list of risk factors identified were for state highways only. See Case Example 2 for details. In summary, a key first step is to identify combinations of high frequency crash location and crash type characteristics that relate to potential systemic roadway treatments and to verify the data needs for the project scope. The target location and crash types then form the basis for the data collection described in Step 2 and a systemic risk analysis described in Step 3. Additional Resources Case Example 1 describes how focus location and crash types were identified for one munici- pal jurisdiction. These additional resources also offer guidance on ways to identify a network and facility/crash type target for more in-depth analysis. Resource Link FHWAâs Guidebook for Developing Pedestrian and Bicycle Performance Measures https://www.fhwa.dot.gov/environment/bicycle_pedestrian/ publications/performance_measures_guidebook/ FHWAâs Guide for Scalable Risk Assessment Methods for Pedestrians and Bicyclists Turner et al., in press FHWAâs Systemic Safety Project Selection Tool https://safety.fhwa.dot.gov/systemic/fhwasa13019/sspst.pdf