SUMMARY

Diagnostic Assessment and Countermeasure Selection: A Toolbox for Traffic Safety Practitioners

Purpose. A Toolbox for Traffic Safety Practitioners provides new tools for diagnosing the contributing factors to crashes and then selecting appropriate countermeasures. It addresses a wide variety of contributing factors to crashes to further practitioners’ understanding of how to balance trade-off decisions that must be made concerning both diagnostic assessments and countermeasure selection. Key goals of this toolbox are to (1) support robust and comprehensive assessments of crashes and subsequent implementation of effective countermeasures and (2) advance the design and operation of roadways that are predictable, support the visibility and conspicuity of key elements, avoid strong demands, and provide all road users enough time to react to the full range of driving situations and conditions.

How to use this toolbox. This toolbox is intended to help those who diagnose contributing crash factors identify and select effective countermeasures for these crashes. The reader can use Chapters 2–9 of this toolbox for search-and-find activities to obtain (1) background knowledge that can further their understanding of key concepts and related research and concepts and (2) practical tools that identify key concepts and questions that can frame and guide the diagnostics process. Chapter 10 contains step-by-step decision trees to aid in countermeasure selection for a broad range of facility types and crash types. Chapter 11 provides a set of techniques and tools that can help practitioners assess roadway design and operations and identify elements that impose a high demand on drivers. Figures, tables, and examples are shared throughout the text to illustrate concepts, present data, and provide templates for practitioner use. Some of these materials appear in more than one chapter to facilitate usage of the varied tools.

This toolbox is intended for use by a variety of traffic safety practitioners, including planners, designers, engineers, and safety analysts. Planners and designers can use it to make initial estimates of the demands that a future facility might place on road users, in terms of high-level topics like meeting expectations and visibility (e.g., sight distance and planned lighting), as well as topics like the complexity of traffic control devices (i.e., workload). Engineers can use it to assess these topics from an operations perspective, as well as to assess more detailed topics like available perception-response times based on features such as signal timing and use scenarios involving mixed modalities (e.g., motor vehicles and bicycles). Safety analysts can use it to assess actual safety performance (i.e., crash frequency and severity), including enforcement records (e.g., speeding violations), tort claims (i.e., evaluations of individual crashes), and groups of similar crashes at the same roadway location (i.e., “hot spots”).

The role of human error in roadway crashes. Several broad, in-depth studies (e.g., Treat et al., 1979; Wierwille et al., 2002; Singh, 2015, and Dong and Wood, 2023) have examined roadway crashes and have found that some form of human error is a key contributor to

most crashes. The types of errors humans commit vary; they can include failing to perceive and recognize a hazardous situation, making bad decisions, or making an error in driver performance. All road users have a limited capacity to receive and process information from the environment; thus, inattention and distraction are frequent contributors to crashes. Also important to a proper understanding of the nature of crashes is to avoid attributing most crashes to a single contributing factor—crashes often have more than one contributing factor (Dingus et al., 2006). An effective framework and methodology for the diagnostic assessment of crashes should include not just a review and analysis of relevant road user, environmental, and vehicle factors, but also the interactions among these factors. Chapter 2 includes a broader discussion of the contributing factors to roadway crashes.

Driver errors can reflect aberrant driver behaviors or human factors issues. While most crashes reflect some sort of driver error, some of these errors (approximately 57%) reflect aberrant driver behavior issues such as impaired driving because of drugs or alcohol, road rage, fatigue, or distraction/inattention, while others (approximately 30%) reflect human factors issues, that is, roadway designs and traffic operation features that place demands on road users that may exceed their capabilities. This distinction between human factors issues and aberrant driver behavior issues is crucial, as they reflect different contributing factors to crashes and corresponding differences in effective countermeasures. Human factors issues include contributing factors to crashes that generally reflect mismatches between the demands placed on the road user by roadway design and traffic engineering features and the inherent physical, perceptual, and cognitive capabilities and limitations of road users. Thus, crashes related to human factors issues may be addressed through some change in the roadway environment (Campbell et al., 2012).

Aberrant driver behavior issues, however, include contributing factors to crashes that generally reflect deliberate violations of law or unsafe driving practices, such as texting while driving, inattention, or driving while impaired by alcohol. Such issues are generally best addressed through behavioral strategies such as training, regulations, or enforcement (Venkatraman et al., 2021). Chapter 4 summarizes research relevant to this topic and provides a diagnostic process to help practitioners distinguish between crashes that primarily reflect human factors issues from those that reflect aberrant driver behaviors.

Roadways are a communications device and are always communicating to road users. A helpful principle provided by the positive guidance approach to design (Lunenfeld and Alexander, 1990; Russell, 1998) is for designers and engineers to consider the highway system as a holistic source of information—a real-time communications device—that is continuously sampled by road users for meaningful information. Critically, roadways are always sending messages to the road user; the key for the practitioner is to design, operate, and continuously evaluate the roadway so that the right messages are consistently provided at the right time to support timely and effective road user behaviors.

A framework for thinking about crashes: roadway demands versus road-user capabilities. Whether human factors issues, aberrant driver behaviors, or some other factors are the culprit, many crashes reflect a fairly simple calculation: crashes are more likely to happen when the demands of the roadway environment exceed the capabilities of the roadway user. In particular, driving requires a complex series of physical actions and mental operations that vary considerably across driving contexts, situations, and conditions. While different driving situations and conditions place different demands on the driver (e.g., making a left turn at an unsignalized intersection against traffic versus driving on an interstate in light traffic), all driving situations require that drivers remain alert and attentive.

Trade-offs in a multimodal transportation network. Trade-offs refer to the inevitable give and take around balancing multiple safety options, including assessments of the strengths and

weaknesses associated with safety considerations for all road users; that is, drivers, pedestrians, bicyclists, and transit users. Both the diagnostic assessment process and the countermeasure selection process should include the compromises, balance, and perhaps exchange between desirable but incompatible elements that characterize making decisions about roadway safety. Diagnostic assessment includes a recognition that road users themselves make trade-offs all the time between convenience, safety, travel time, and costs and that crash assessments should take this into consideration. Countermeasure selection by the safety professional should include trade-offs between key variables, including countermeasure efficacy, specific safety benefits, unanticipated outcomes, and feasibility (e.g., time and cost).

Importance of evaluation. Critical to improving safety performance (i.e., crash frequency and severity) is the evaluation (a fifth “E” that can be added to the four “E’s” of highway safety—engineering, education, enforcement, and emergency medical services) of crash data in modal and facility contexts to assess crashes and aid the selection and design of countermeasures. While program evaluation might be considered something to worry about after countermeasures have been identified, this fifth “E” should be implemented at every stage of the safety improvement process and include input and involvement from the range of transportation professionals involved, including planners, designers, engineers, and safety analysts. In short, having an evaluative mindset throughout the crash prevention process can add rigor and purpose to safety improvement planning. Several chapters of this toolbox discuss the importance of evaluation throughout the diagnostic assessment and countermeasure selection process and include tools and methods for conducting evaluations, especially Chapters 3 and 9.

Diagnostic Assessment in the Safe System

Overview of the Safe System approach. The Safe System approach seeks to plan, design, and operate a road system that recognizes that humans make mistakes, have limited physiological abilities to safely negotiate complex situations, and have a limited tolerance of kinetic energy forces (Signor et al., 2018; Welle et al., 2018; Finkel et al., 2020). A key goal of this holistic approach is to create a system that reduces the risk of kinetic energy transfer occurring in the first place and reduces the amount of energy transfer in the event of a crash to an amount that can be tolerated by humans. Overall, the Safe System approach incorporates five elements: safe road users, safe vehicles, safe speeds, safe roads, and post-crash care. The diagnostic process should very much be considered to be an evaluative activity, as the fifth “E” (evaluation) should exist at every stage of the safety management process and not just toward the end of the process as part of countermeasure evaluation. The approach includes the five “E’s” of traffic safety but also encompasses planners, designers, operators, and users of the transportation system to prevent fatal crashes and reduce crash severity.

A general process for diagnostic assessment. A diagnostic assessment process that incorporates the holistic elements associated with Safe System will explicitly incorporate general consideration of all road users but also

• Consider pre-, during-, and post-crash factors that might have contributed to the crash itself, as well as post-crash survivability;
• Evaluate specific issues related to the role of perception-response time, expectations, visibility, and workload/demand; and
• Produce a detailed summary of possible contributing factors and their likely interactions.

Figure 1 summarizes a process for diagnostic assessment and emphasizes the importance of considering not just crash and site data but also key human factors issues (discussed in more detail in this summary and in Chapters 5–8) that are often the key contributing factors

to crashes. It will also be useful to incorporate aberrant driver behaviors (such as impaired or distracted driving) into the analysis.

The focus at this stage of the diagnostic assessment process is on thinking like a virtual road user and identifying and understanding the roadway components that could contribute to confusion, poor visibility, misperceptions, high workload, distraction, or other potential road user errors at a particular site. Solutions in the form of countermeasures or treatments (discussed in Chapters 9 and 10) are not sought at this stage of crash evaluations.

Figure 1 includes a step to develop a Modified Haddon Matrix as part of the diagnostic assessment process. Haddon’s epidemiological view of injury (Haddon, 1972; National Committee for Injury Prevention and Control, 1989) outlines three phases to crash evaluations:

1. A pre-crash phase that includes those factors that influence whether a crash will occur and then result in injuries
2. A crash event phase that includes those factors that influence injury severity during the crash event
3. A post-crash phase that includes those factors that influence the survivability of the crash after the event

To augment this approach, John Milton and Ida van Schalkwyk of the Washington State Department of Transportation (personal communication, January 17, 2022) have developed a framework that directly considers all road users (e.g., the volume of biking and walking) and supports the social safety environment. Consistent with the Safe System approach, it has also added user-mix considerations and interactions between these factors. Table 1 shows a Modified Haddon Matrix applied to motor vehicle crashes in the Safe System. This matrix was developed by Milton and van Schalkwyk and is used with their kind permission.

By introducing social environment factors, safety professionals are asked to consider the implications of attitudes, biases, and equity decision-making frameworks for humans operating in the roadway environment. Doing so expands the potential diagnostic assessments that safety professionals perform.

Table 1. Modified Haddon Matrix applied to motor vehicle crashes in the Safe System.

This modification of the original Haddon Matrix aims to present a framework that more directly considers all road users and the supporting social safety environment. It is best to generate a Modified Haddon Matrix that includes all possible factors impacting safety performance (crash frequency and severity). Specifically, the Matrix should include any factors and combinations of factors (interactions) that could reasonably contribute to the known or suspected opportunities for reducing crash potentials at the site under investigation. Broad considerations should include crash or conflict type, frequency, severity, and contributing factors, as well as on-the-scene observations of the facility and representative traffic movements (including pedestrians, bicyclists, and transit vehicles). In this regard, some interactions may be quantitative and specific, while others may be qualitative and reflect possible impacts.

Chapter 3 presents a more detailed discussion of diagnostic assessment in the Safe System.

Evaluating the role of human factors issues in crash diagnostics. Treat et al. (1979) and several subsequent large-scale crash studies have shown that, while most crashes indeed reflect some form of driver error, a subset of these errors (approximately 57%) reflect primarily “driver-only” issues such as impaired driving because of drugs or alcohol, road issues (including roadway designs) and traffic operation features place demands on road users that may exceed their capabilities (i.e., human factors issues). Human factors issues, then, are simply interactions between the roadway and the road user that contribute to crashes.

While a thorough review of human factors is beyond the scope of this toolbox, decades of crash analyses and studies of driver behavior and performance (Wierwille et al., 2002; Campbell et al., 2012; Theeuwes and van der Horst, 2012; Permanent International Association of Road Congresses [PIARC], 2019a) make clear that the following four human factors issues are key to the efficacy of roadways as communication devices and the general safety performance (i.e., crash frequency and severity) of roadways:

Perception-response time (PRT)—A common thread in the circumstances that lead to many roadway crashes is time, specifically, a lack of time for the driver (for example) to respond to an upcoming object in the roadway, a change in roadway characteristics or a navigation demand. PRT is comprised of multiple components that generally overlap in time and includes the time to detect a target, process the information, decide on a response, and initiate a reaction. The other key human factors issues—expectations, visibility, and workload—primarily present challenges to a driver because of the time constraints imposed by the driving task. A large body of research involving laboratory and on-road studies has helped estimate approximate driver PRT values under a variety of situations. Although this range is wide (roughly 0.5 to 2.5 seconds), the data on PRTs can have significant utility in crash diagnostics and roadway design. The “Green Book” (American Association of State Highway and Transportation Officials [AASHTO], 2018) recommends a design criterion of 2.5 seconds to include the capabilities of most drivers under most highway conditions, but actual PRTs will be situation dependent. Specifically, driver PRTs will be determined by several factors, including driver expectations, driver age, the conspicuity of detected roadway objects, and vehicle speed.

In a crash sequence, vehicle speed translates directly into the time available for drivers to react; the time issue therefore highlights the importance of posted speed selection by designers and the speed choices made by drivers. Perhaps one of the most direct ways that roadway designers can improve safety performance (i.e., crash frequency and severity) is to design the roadway and accompanying operations to give more time for road users to perceive, decide on a response to, and react to situations and conditions.

Expectations—Drivers rely on their general knowledge and past experience driving a particular facility to aid the continuous driving task and manage new information they need to process. Designing roadway environments in accordance with driver expectations

is a crucial way to accommodate a road user’s inherent human limitations and make the roadway predictable for road users. The idea of predictability in design is key, as basic facility design and related traffic operations should avoid surprising road users. Predictability is the key benefit to meeting users’ expectations; predictability allows the road user to place a roadway into a predefined category that—in turn—dictates behavior. In this regard, roadways that meet user expectations can function like a familiar “script” that directs and supports behaviors that are both effective and timely. Navigating the roadways as a driver, bicyclist, or pedestrian benefits from physical environments, sequences of required actions, and interactions with others that are predictable. Expectations are central to how a driver monitors, perceives, and interprets the roadway environment, makes decisions, and then acts or responds. Variances from driver expectations are a key source of misperceived and misinterpreted information, slowed response times, and driver errors.

For the driver, expectations are closely related to the broader principle of design consistency. Design consistency is one of the most basic principles in human factors and system design. Consistency improves user performance because it facilitates the user’s ability to predict what the system will do in any given situation; it provides rules that govern relationships between elements in the environment (e.g., signage and geometry) and driver responses (e.g., navigation and speed selection). The same underlying principle applies to other road users, such as bicyclists, pedestrians, and transit users.

Visibility—Visibility refers to the quality or state of being visible and relates to concepts such as conspicuity and sight distance. Approximately 90% of the information used by a driver is obtained visually (Hills, 1980) and, according to Treat et al. (1979), one of the leading environmental factors related to traffic crashes is visibility. Visibility is closely related to conspicuity, defined by Krauss (2015) as “those characteristics of an object or condition that determine the likelihood that it will come to the attention of an observer” (p. 57). In other words, conspicuity refers to how noticeable or visible an object is.

A driver’s ability to select and act on even highly visible and conspicuous objects may be impacted by the amount and type of competing stimuli (e.g., pedestrians, other vehicles, signs, markings, buildings, or planters) in the immediate environment (Edquist and Johnston, 2008). Thus, the complexity and clutter associated with a visual scene impact visibility. Drivers’ visibility may be reduced because of physical limitations, such as inclement weather conditions (e.g., fog, rain, snow, or sun glare), vehicle conditions (e.g., burned-out headlights or a dirty windshield), the geometry of the roadway, and other scene elements. Regarding perceptual or cognitive limitations, visibility may also be impacted by driver distraction or inattention, diminished visual acuity, or reduced contrast sensitivity (Hills, 1980).

In general, road signs, lane markings, and lighting are roadway-design elements that can significantly impact visibility, especially at night (FHWA, 2022a), and help address visibility limitations based on roadway geometry (e.g., sharp horizontal curves or vertical curves). For example, advance signage can inform the driver of upcoming limited visibility situations, and well-maintained retroreflective signs can improve nighttime visibility and reduce the risk of crashes by making signs appear brighter and easier to see and read from a distance. Similarly, well-maintained lane markings can make road curvatures more visible from a greater distance, thereby preparing the driver for unexpected curves. Finally, roadway lighting helps make roadways, surrounding areas, and road users, such as pedestrians, more visible at night.

Task demand/Workload—Workload refers to the overall demands placed on an individual by a particular activity, including effort, task complexity, time usage, and the nature of possible interference between concurrent tasks (Gartner and Murphy, 1979; Gawron, 2000; Angell et al., 2006). The workload in the driving context generally refers to the demands

placed on an individual while performing driving tasks, which can be impacted by driver effort, task complexity, time constraints, and potential interference among multiple concurrent tasks (Tijerina et al., 1996). A useful shorthand for thinking about workload is to consider the time available to complete a task relative to the specific demands of the task; in other words, workload = time/task. While effort and time pressures are key aspects of workload, in the driving domain, workload is frequently described in terms of the following (Richard et al., 2006):

• Perceptual requirements (e.g., detecting and making sense of what is seen, heard, and felt)
• Decision-making/cognitive needs (e.g., making decisions to go, stop, or turn; integrating what you perceive with things you know, such as rules of the road; or using your previous experience with a roadway facility to help decide what maneuver is appropriate)
• Psychomotor/response needs [e.g., executing a decision, taking an action (such as a driver changing their point of gaze to somewhere else within the visual scene), braking, or changing lanes]

Driving is a demanding activity, as drivers are continuously performing several tasks at the same time, often while moving at a high rate of speed. These tasks include visual scanning of the environment, perception and identification of task-relevant elements (e.g., other vehicles, pedestrians, traffic signs, and traffic signals), lane maintenance, and speed control. Critically, the driving environment is ever-changing, often requiring relatively high vigilance and attention to these tasks. Importantly, workload concerns are not limited to drivers, and the demands placed on other road users should be considered as well. For example, pedestrians crossing a roadway need to plan a path, monitor for hazards such as vehicles and bicycles, and then execute a crossing maneuver. Likewise, bicyclists riding within the lane or on the shoulder of a rural road have to watch the roadway ahead, maintain position within a lane or bike path, and continuously monitor for and avoid in-path objects.

The nature and source of high task demands will vary considerably across different roadway types and driving conditions. Table 2 shows some typical roadway types and highlights features that could increase demands.

Chapters 5 through 8 in the toolbox provide more comprehensive discussions of these four human factors issues, including questions for each that can be used as part of the diagnostic assessment process presented in Figure 1. Quantifying workload/demand for drivers requires conducting workload analysis for a given roadway segment and driving task, and Chapter 11 provides a set of techniques and tools that can help practitioners assess roadway design and operations and identify elements that impose a high demand on drivers. These methods aim to consider what a roadway requires of drivers and include ways to assess the number of static and dynamic elements that require the driver’s attention, comprehension, decisions, and potential action, as well as more involved workload assessment techniques.

Selecting Effective Countermeasures

Effective countermeasures decrease the demands placed on the road user and/or augment the road user’s capabilities in some fashion. Roadways are communication devices, and they are at all times communicating a host of messages to the road user. Sometimes these messages are helpful and intended by the roadway designer, and sometimes not. Countermeasures provide an opportunity to improve how the roadway designer communicates with the road user and can decrease the crash frequency or severity at a particular site. Once again, many crashes happen when the demands of the roadway environment exceed the capabilities of the roadway user. In this regard, it is important to recognize that roadway countermeasures

Table 2. Roadway types and elements that can increase demands on road users (photos by Samuel Tignor, 2021; used with permission).

Roadway Types	Elements That Can Increase Demands on Road Users
Interstate Highways	High speeds can translate into reduced time to respond to conflicting vehicles. High levels of visual demand, including frequent glances to adjacent lanes, are required to respond to multi-vehicle moving hazards (e.g., merge maneuvers, cut-ins, and stopped vehicles ahead). Conflicting moving vehicles could enter the road from the ramp. Perceptual and decision-making demand is higher with smaller gaps (dense traffic) for merging and lane changing.
State Routes	Curves and vegetation obstruct delineation cues and oncoming vehicles near a hill. Potential for hidden driveways Visibility/expectancy at intersections Presence of bicyclists and pedestrians Visibility/expectancy is questionable at unexpected intersections. Possible conflicts at the crest of a hill and turning road Guardrail close to roadway restricts possible escape path.
Rural Roads	Potential vehicles, pedestrians, or bikes at or beyond the crest Narrow lane widths on each side of the road Potential for hidden driveways Narrow roadway with no centerline plus narrow shoulders An unexpected vehicle can enter the road from a driveway.
Urban	Highly dynamic traffic elements–stopping and turning vehicles, many bicycles, scooters, and pedestrians–are cramped into the intersection. Challenging road user guidance and navigation in a dense-visual environment Visibility of other road users (e.g., the two bike lanes in the center of the street are hard to perceive, and pedestrians may be blocked by other vehicles). Two-way traffic on a cross street Traffic signals and signs must be perceived and comprehended.
City Center	A mix of road users—pedestrians, bicyclists, scooters—often in unexpected locations Many traffic control devices—traffic signals, signs Dynamic traffic elements—stopping and turning vehicles Challenging road user guidance and navigation in a dense-visual environment Extremely limited visibility/expectancy throughout the street Large trucks and buses block users’ visibility and mobility. Traffic signals, signs, and pavement markings are blocked by queued vehicles. User workload is extremely high for all modal users.

operate within this demand versus capabilities framework. Specifically, effective countermeasures decrease the demands placed on the road user and/or augment the road user’s capabilities in some fashion. Accordingly, countermeasures should be selected in a manner that links their features and benefits to the underlying contributing factors and human factors issues observed within the crash data or the facility itself.

A countermeasure selection process. This chapter’s discussion regarding the countermeasure identification and selection process picks up where the diagnostic assessment process (as illustrated in Figure 1) ended—with a partially completed Modified Haddon Matrix. Figure 2 illustrates the countermeasure identification and selection process and is primarily intended to emphasize the importance of linking countermeasure identification and selection to specific human factors issues (as well as driver behavior issues) identified during the diagnostics process.

The features and benefits of specific countermeasures should be matched to the underlying crash-contributing factors observed within the facility. As an illustration of the need to link countermeasure features with underlying contributing factors to crashes, consider the growing issue of pedestrian fatalities. In 2021, there were 7,388 pedestrians killed in traffic crashes; this reflects a 12.5% increase from 6,565 pedestrian fatalities in 2020 (National Center for Statistics and Analysis [NCSA], 2023a). In 2021, pedestrians accounted for 17% of United States fatalities, up from 14% in 2012. While alcohol is involved in many pedestrian crashes, other characteristics of fatal pedestrian crashes are more closely aligned with human factors issues. For example, 84% occurred in urban areas, 75% occurred outside of intersections, and 77% occurred in the dark. While visibility issues seem obvious from the percentage of pedestrian fatalities occurring in the dark, the fatalities occurring in “outside intersections” locations suggest impacts of expectations—drivers are more likely to anticipate pedestrians near intersections.

Several technology development efforts and safety evaluations have been conducted to help address crashes and fatalities involving pedestrians.

Table 3. Example countermeasures for pedestrian crashes and how they help road users.

Table 3 (adapted from Brown et al., 2021) highlights some of these countermeasures and explains how they help road users for example, by reducing speeds (thus giving drivers more time to perceive and react to pedestrians) and/or increasing pedestrian visibility and conspicuity.

Decision trees for selecting countermeasures. Chapter 10 of this toolbox provides a series of decision trees to help practitioners select countermeasures to address target crash types and facility types. The decision trees provide a visual framework for decision-making. Practitioners can select from a series of decision trees that lead them through diagnostic questions to help identify countermeasures that could potentially address crash-contributing factors associated with the crash pattern of interest. Selecting countermeasures for potential implementation, matched to underlying contributing factors to target crash types, is expected to reduce crashes to the greatest extent possible.

The approximately 80 distinct decision trees in Chapter 10 address common crash types that occur along rural and urban roadway segments and intersections, including crashes within the following:

• Rural two-lane roadway segments
• Rural multilane undivided roadway segments

• Rural multilane divided roadway segments
• Urban two-lane roadway segments
• Urban multilane undivided roadway segments
• Urban multilane divided roadway segments
• Rural and urban signalized intersections
• Rural and urban unsignalized intersections

The crash types are further categorized according to common crash contributing factors. Through the diagnostic process outlined in Chapter 3 of this toolbox, practitioners should get a sense of the factors contributing to the crash pattern of interest. Then, by reviewing the respective decision trees, following the logic, and responding to the questions, practitioners will be able to identify a list of potential countermeasures for further consideration in the economic appraisal and project prioritization process. The order of countermeasures presented from top to bottom in the decision trees is not intended to signify any type of prioritization for countermeasure selection or implementation. Priority for selection and implementation is to be based on the applicability of the countermeasure to remedy the crash type of interest at the given site, the cost effectiveness of the implementation of the countermeasure, and other factors, such as trade-offs in safety and mobility between all road users.

Prioritizing candidate countermeasures. The user may want to give a countermeasure greater consideration for implementation if (1) the countermeasure was identified for implementation more than once in response to different diagnostic questions or (2) when the same countermeasure was identified for potential implementation in response to different combinations of contexts, crash types, and contributing factors. Even in these situations, the economic appraisal and project prioritization process, following countermeasure selection, will provide additional details to inform decisions regarding prioritization for implementation based on economic performance measures.

Sample decision tree. Figure 3 shows an example of a decision tree from Chapter 10. In this decision tree, the driving context is rural two-lane segments in which road departure crashes (single-vehicle run-off-road/head-on/sideswipe, opposite direction) are occurring under superelevation and possible decrements in the roadway surface conditions. Critically, final countermeasure selection should include trade-offs between key variables, including countermeasure efficacy, specific safety benefits, unanticipated outcomes, and feasibility (e.g., time and cost).