APPENDIX A

Style and Usage Guide of Terms and Phrases to Be Used and Avoided in HSM2

Introduction

This Style and Usage Guide of Terms and Phrases to Be Used and Avoided in HSM2 was intended to guide the research team as it drafted and revised content in the production of the second edition of the Highway Safety Manual (HSM2). During the development of chapter content and review process, other terms and phrases that came to light were incorporated into this style and usage guide.

Preface

The HSM is not a manual of standards and warrants, nor is it a guidance document. Instead, it is a resource that provides analytical tools, techniques, and procedures for quantifying the potential effects of decisions made by transportation professionals. It is a toolbox for practitioners and researchers.

It is important that the HSM2 not inadvertently create tort liability implications in its content. Carelessly drafted language and the use of certain terms and phrases can undermine the utility of a document by allowing for misinterpretation or misuse by those in litigation against transportation agencies.

This style and usage guide includes examples and explanatory material that inform the research team of the meaning and effect of certain terms and phrases that carry legal implications. Content that is subjective and based on personal opinion is to be avoided. Instead, content based on objective, specific, and measurable data is to be used. Vague or ambiguous content subject to misinterpretation or multiple interpretations is to be made more precise. Directive or mandatory language that leaves no room for the exercise of engineering judgment is to be eliminated or at least minimized. Content that sets the user up for failure by establishing performance levels beyond that which can be reasonably achieved is to be made realistic. Finally, terms and phrases with pejorative and negative implications that lend themselves to admissions of wrongdoing or liability are to be explained and neutral objective language offered as alternatives.

The context of these terms and phrases determines whether they clearly state what is intended or if they have the potential for unintended consequences. To aid the research team, this style and usage guide approaches its task by categorizing the terms and phrases into particular contexts to more fully illustrate why and how they should be avoided or used:

• Subjective terms and phrases expressing personal opinion.
• Vague and ambiguous terms and phrases.
• Terms and phrases that are directive or mandatory.
• Terms and phrases setting up for failure.
• Terms and phrases that are pejorative and have liability implications.

Subjective Terms and Phrases Expressing Personal Opinion

The following terms and phrases are used to express a relatively positive or negative opinion. They can be lazy substitutes for data-driven conclusions. The writer attempts to be omniscient but in reality is only offering the writer’s personal opinion, usually without giving the user any specific support for it. They purport to reach a conclusion about what the data demonstrate without sharing the underlying data with the user. To some degree, they tend to insult the user by saying: “Trust me. I know better than you. I don’t need to support my opinion.”

• Adequate/sufficient. These terms are used by the writer to quantify without being precise. They have a positive implication and are presumably linked to some subjective scale in the writer’s mind. What is being measured meets a certain threshold in the writer’s unsupported opinion, thereby moving that which is being measured out of the “inadequate” or “insufficient” category, whatever that is. The material in the HSM2 is quantitative in nature and is expressed in data form. The user should be given the opportunity to evaluate those data and determine where they fall on an objective quantifiable scale.
• Better/best. These terms express the writer’s opinion as to relative positivity. [e.g., “The ‘better’ (or ‘best’) approach for applying CMFs to estimate crash frequency is methodology A”]. They are used without presenting any objective data to support the conclusion. Since the HSM2 is a toolbox and not a guidance document, it is more useful to display the data and then allow users to reach a conclusion as to which is more applicable to the issue before them. (e.g., “From the foregoing data, one can determine that methodology A produces the most reliable result.”)
• Deficient. This term is used by the writer to quantify without being precise. (e.g., “The stopping sight distance was ‘deficient’.”) It has a negative implication and is presumably linked to some subjective scale in the writer’s mind. Moreover, it is not merely that what is being measured fails to meet a certain value in the writer’s unsupported opinion; it fails to meet that value by an amount that undercuts its utility. The material in the HSM2 is quantitative in nature and is expressed in data form. The user should be given the opportunity to evaluate the data and determine where they fall on an objective quantifiable scale. (e.g., “A stopping sight distance of 45 ft. is lower than that recommended in the AASHTO Green Book.”)
• Desired/desirable. These terms express the writer’s opinion as to relative positivity. (e.g., “This formula will produce the ‘desired’ result. It is ‘desirable’ to examine crash rates both before and after an improvement.”) They are used without presenting any objective data to support the conclusion. Since the HSM2 is a toolbox and not a guidance document, it is more useful to display the data and then allow users to reach a conclusion as to which is more applicable to the issue before them. (e.g., “Examining crash rates both before and after an improvement provides information on the effect of the improvement on crash rates.”)
• Difficult. This term usually expresses the writer’s opinion about a task or procedure and rarely is accompanied with supporting data. Whether a methodology is “difficult” to use or a stop sign is “difficult” to see is for the user to determine based on the data. It is presumptuous for writers to impose their unsupported opinions on the user.
• Excessive. This term is used by the writer to quantify without being precise. It has a negative implication and is presumably linked to some subjective scale in the writer’s mind. Moreover, it is not merely that what is being measured exceeds a certain value in the writer’s unsupported opinion; it exceeds that value by an amount that moves it into the qualitative “excessive” category, whatever that is. (e.g., “The number of crashes at the site was ‘excessive’.”) The material in the HSM2 is quantitative in nature and is expressed in data form. The user should be given the opportunity to evaluate those data and determine where they fall on an objective quantifiable scale. (e.g., “The number of crashes at the site exceeds that expected on average for such a site.”)
• Inadequate/insufficient. These terms are used by the writer to quantify without being precise. They have a negative implication and are presumably linked to some subjective scale in the

• writer’s mind. Moreover, it is not merely that what is being measured fails to meet a certain value in the writer’s unsupported opinion; it fails to meet that value by an amount that moves it into the qualitative “inadequate” or “insufficient” category, whatever that is. The material in the HSM2 is quantitative in nature and is expressed in data form. The user should be given the opportunity to evaluate those data and determine where they fall on an objective quantifiable scale.
• Just. The term “just” is a qualifier. It implies an inchoate insufficiency. (e.g., “The shoulder was ‘just’ 3-feet wide.”) It expresses a personal opinion without any data to back it up. Objective measurements, data, and assertions are not enhanced by the use of such an unsupported opinion.
• Poor. This term expresses the writer’s negative opinion. (e.g., “The roadway is in a ‘poor’ condition.”) The term is used without presenting any objective data to support the conclusion. It also expresses that opinion in a pejorative way by denigrating the subject of the opinion. Since the HSM2 is a toolbox and not a guidance document, it is more useful to display the data and then allow users to reach a conclusion as to which is more applicable to the issue before them.
• Preferred/preferable. These terms express the writer’s opinion as to relative positivity. (e.g., “This methodology is ‘preferred.’ It is ‘preferable’ to use at least 3 years of data.”) The terms are used without presenting any objective data to support the conclusion. Since the HSM2 is a toolbox and not a guidance document, it is more useful to display the data that will allow users to reach a conclusion as to which is more applicable to the issue before them. (e.g., “Using at least 3 years of data yields a more reliable result.”)
• Safe/unsafe. These terms may express the writer’s subjective opinion of the safety performance of a roadway, feature, condition, or facility. However, they also may express the writer’s subjective opinion of how secure that roadway, feature, condition, or facility makes them feel. This feeling is quite apart from the objective frequency and severity of crashes and reflects the subjective reaction of the writer. As a result, it has no utility to the user of the HSM2. The users need data on which to base a conclusion.
• Worse/worst. These terms express the writer’s opinion as to relative negativity. (e.g., “This intersection is ‘worse’ than that one. Option A is the ‘worst’ option.”) The terms are used without presenting any objective data to support the conclusion. The terms also express that opinion in a pejorative way by denigrating the subject of the opinion. Since the HSM2 is a toolbox and not a guidance document, it is more useful to display the data and then allow users to reach a conclusion as to which is more applicable to the issue before them.

Vague and Ambiguous Terms and Phrases

The following terms and phrases suffer from either having no precise meaning or clear direction of what is intended (are vague) or having more than one meaning of what is intended or directed (are ambiguous). These terms are sometimes used as a default when the writer is uncertain. They are an attempt to highlight an issue without explaining how it may be resolved.

• Care should be taken. This phrase, as well as “use with caution,” is intended to act as a warning to users that the subject matter to follow has some limitation. Unfortunately, it does not advise the user as to what the specific limitation is, nor does it provide any data to illustrate how the limitation may affect a result or outcome. The writer needs to clearly state what the precise limitation may be and what can be done to overcome that limitation.
• Often. This term is a vague expression of the frequency of an event. The user can assume the frequency is more than once but, beyond that, one has no clue. Since it is a term used as a substitute for data, the user is left to guess its significance or whether it has any significance at all. If frequency is an issue to be discussed, set forth the data and state the frequency with precision.
• Problematic. This term has a negative connotation and is used to alert the user that something is amiss, but the user has no idea what it may be. It may mean the result is questionable, implausible, dubious, puzzling, or uncertain. It does tell the user that something is suspect, but

• without more, the user does not know how to act on the information. It is only useful information if the issue is specifically described, and the user is advised what can be done about it.
• Questionable. Like “problematic,” this term has a negative connotation. It implies that something is wrong with a methodology or result. It does not tell the user the magnitude of the issue or whether it can be overcome. Without some specificity on its limitations, the user cannot determine whether the “questionable” information is useful or not. It is a descriptor to avoid in the HSM2.
• Safe. The term “safe” can refer to different concepts. “Safe” can describe how a person feels in a situation—whether a person feels threatened or in harm’s way. It can also refer to a facility or roadway at a particular site in terms of the experienced or expected frequency and severity of crashes—something that can be quantitively measured. The HSM2 is data-driven, and vague or ambiguous terms such as “safe” should be replaced with ones that are measurable and based on specific data.
• Safety. The term “safety” has been clearly defined in the HSM to refer to the “frequency and severity of crashes” (AASHTO, 2010). In that context, the use of the word in terms such as “Highway Safety Manual,” “road safety principles,” “safety performance functions,” and “safety analysis” is perfectly acceptable shorthand for that concept. On the other hand, the term “safety” can be overused and become lazy shorthand. Sometimes a more specific reference to the “frequency and severity of crashes” is much clearer. Whether “safety” is an appropriate term to use will depend on the context.
• Troubling/troublesome. These terms have a negative connotation and attempt to vaguely alert the user to an issue without specificity. Like “problematic,” these terms tell the user that something is suspect or worrisome, but without more, the user does not know how to act on the information. It is only useful information if the issue is specifically described, and the user is advised what can be done about it.
• Use with caution. This phrase, like “care should be taken,” is intended to act as a warning to users that the subject matter to follow has some limitation. Unfortunately, it does not advise the user as to what the specific limitation is, nor does it provide any data to illustrate how the limitation may affect a result or outcome. The writer needs to clearly state what the precise limitation may be and what can be done to overcome its limitation.

Terms and Phrases That Are Directive or Mandatory

Unlike terms and phrases that are vague or ambiguous, directive or mandatory terms and phrases clearly tell the user what to do. They do not allow for a variance, using discretion, or engineering judgment. Deviation from these dictates can leave a practitioner exposed to liability for failure to meet a standard of care or worse, for a violation of law.

• Essential/imperative/necessary. These terms express the importance of certain actions without any objective data to support them. They do not allow for discretion or engineering judgment. They presuppose that someone has already decided there is no room for independent decision-making. Failure to comply may be equated to a failure to meet a standard of care. Given that the HSM2 is a toolbox and not a guidance document, these terms are to be avoided, and instead, users are to be given the data that supports the importance of an action. (e.g., “The foregoing crash data illustrate that the site may be a candidate for an improvement.”) Users then make the decision and are not simply doing what they are told to do.
• Guide/guidance. While this style and usage guide is indeed a guidance document to be used in drafting and revising content for the production of the second edition of the Highway Safety Manual, the HSM2 itself is not a guidance document. It does provide information about analytical tools, techniques, and procedures for quantifying the potential effects of agency decisions on the frequency and severity of crashes. What it does not do is recommend a particular

• result, decision, outcome or course of action based on that information. Therefore, “guide” and “guidance” are terms to be avoided in describing the content of the HSM2.
• Mandatory. This term speaks for itself. It is a command. It leaves no room for discretion or judgment. The action to which it refers must be done unqualifiedly. If not, the user will have failed to meet a standard of care and may even be in violation of law. The term “mandatory” should not be found in the HSM2.
• Must. This term has a mandatory connotation. In the context of telling the user that a certain improvement must be implemented, it suffers from the same liability implications that the word “shall” does: the user may be exposed to liability for failure to implement the improvement. On the other hand, when used in the context of explaining the correct methodology to use in a given situation (e.g., “Step 1 ‘must’ be completed before going on to Step 2”), “must” can be an appropriate word to use. While the HSM2 is not a guidance document, it does contain instructions on the use of various tools and informs the user on how to use those tools.
• Need/needs. These terms connote a certain urgency for some action without any measurable data to support it. (e.g., “The existing shoulder ‘needs’ to be widened to 10 ft.”) It limits discretion and engineering judgment. In the context of whether an improvement is to be implemented, the user is better served by suggesting that the location “may be a candidate for an improvement” or “may benefit from an improvement.”
• No less than/no more than. While these terms imply that there is some room for discretion or engineering judgment, they also connote that there is a limit beyond which they become mandatory. (e.g., “Shoulders are to be ‘no less than’ 8 ft. wide.”) A toolbox like the HSM2 allows for discretion, and its language should reflect that. (e.g., “Shoulders are to be 8 ft. wide where feasible and subject to engineering judgment.”)
• Regulation. A regulation is a rule of law and has the force and effect of law. Violation of a regulation is a violation of law, which subjects a user to independent liability. The HSM2 does not contain regulations, and the term should not be used.
• Required/requirement. These terms are mandatory. There is no room for discretion or engineering judgment. (e.g., “The stop signal is ‘required’ at this intersection.”) A failure to do that which is required or a failure to meet a requirement falls short of meeting a standard of care. This can leave the user subject to liability. A phrase such as “may be considered” allows the user to examine the data and then exercise discretion and engineering judgment.
• Shall. The term “shall” tells the user what to do, usually in the context of a manual, guideline, or regulation. It describes an action that is mandatory or required and leaves no room for the exercise of discretion or engineering judgment. When used in a regulation, it has the force of law and creates a mandatory duty to comply. Failure to comply in that situation creates an independent cause of action for violation of a mandatory duty and subjects the user to liability. If “shall” is used in a manual or guideline that does not have the force of law, it may nevertheless establish the standard of care in an action for negligence. If users do not meet that standard of care, they become subject to liability. The term “shall” is only to be used with full knowledge of its consequences and when the action must be taken without the exercise of discretion or engineering judgment. Since the HSM2 is neither a set of standards nor a guidance document, the term “shall” is not to be used to direct the user to act. Permissive terms such as “may,” “can,” or “could” provide the broadest discretion, allow for a variance based on sound engineering judgment, and are less likely to expose the user to liability.
• Should. The term “should” advises the user to take a recommended course of action under particular circumstances. While it does allow for some discretion and the exercise of engineering judgment, it carries with it the implication that deviation from its recommendation is discouraged. Since the HSM2 is neither a set of standards nor a guidance document, the user is best served by providing the data on which an informed decision can be made. Permissive terms such as “may,” “can,” or “could” provide the broadest discretion, allow for a variance based on engineering judgment, and are less likely to expose the user to liability.

• Standard. A standard is commonly a statement of a required, mandatory, or specifically prohibitive practice. Deviation from a standard brings with it possible exposure to liability for not meeting a legal standard of care. Standards can be found in guidance documents but not in a toolbox such as the HSM2.

Terms and Phrases Setting up for Failure

These terms and phrases set the user up for failure by establishing performance levels beyond those that can be reasonably achieved. Sometimes they may be well-intentioned aspirational goals. However aspirational they may be, they are still goals that cannot be met. Moreover, to the extent these goals are adopted by an organization, not attaining them is an admission of failure. Such an admission can have liability implications in a lawsuit.

• All. The term “all” can be an over-inclusive term. Designing a facility that is accessible to “all” users is an unlimited concept. What is encompassed by the word “all” can vary over time and depend on the circumstances. It establishes a bar too high to be achieved, setting one up for failure. In this context, state with particularity the group that is to be included (e.g., motorists, bicyclists, and pedestrians).
• Always. When the term “always” is used to predict an outcome, it is expressing a certainty. (e.g., “Alternative treatment A ‘always’ results in a reduction of the frequency and severity of crashes.”) However, crash modification factors, for example, have some standard error attached. By definition, there is some degree of unpredictability in an outcome. Failure to account for that can set the user up for failure. In this context, the word “always” is to be avoided.
• Ensure. The term “ensure” does not allow for exceptions. It imposes an overwhelming duty on the actor that has legal implications if that duty is not met. (e.g., “The user must ‘ensure’ that all measurements made in the field are correct before this methodology can be used.” This would require at a minimum that a user personally go out in the field and verify each measurement.) It is a burden that cannot reasonably be met and sets the user up for failure.
• Possible. Almost anything is “possible” given the time, resources, materials, tools, expertise, ingenuity, science, and funds. Thus, when users are instructed to do something “as soon as possible,” “where possible,” or “when possible,” they are being set up for failure. Simply put, all possibilities cannot be anticipated. To allow for discretion and engineering judgment, the words “practical,” “practicable,” or “reasonably feasible” should be used instead.
• Will. When the term “will” is used without a qualifier to predict an outcome, it is expressing a certainty. (e.g., “Alternative treatment A ‘will’ reduce crashes.”) However, all crash modification factors, for example, have some standard error attached. By definition, there is some degree of unpredictability in an outcome. Failure to account for that can set the user up for failure. This can be avoided when the word “will” is qualified by “on average,” “in many cases,” and so forth.

Terms and Phrases That Are Pejorative and Have Liability Implications

Pejorative terms and phrases that have liability implications are of particular interest to litigants and their attorneys. They can be terms that have a common meaning to the user, yet at the same time, they may have a legal meaning or connotation as well. They can cast an unintended judgment or call for immediate action without conscious awareness of the legal consequences. In the hands of a skilled plaintiff’s attorney, they can be used against you.

• A concern/is concerned/of concern. These terms are not neutral. To have a concern about something, to be concerned about something, or to have something be of concern implies that something is bad. There is a problem about which to be concerned. This is to be distinguished from

• “business concerns” (i.e., interests) and matters concerning data sampling (i.e., about, related to). Using “concern” in the former context exhibits subjectivity. Whatever the issue is, there is already a conclusion that the issue has a negative impact and needs attention. Having a concern also implies having notice of a problem. Once an agency has notice, the clock is ticking for an immediate solution. Failure to act can give rise to liability. Using terms that are more neutral or even positive, such as “matters,” “issues,” or “opportunities,” can be more specific and objective.
• Danger/dangerous. Aside from suffering from an extremely pejorative connotation, these terms are also legal words of art; that is to say, they have a legal meaning in addition to their normal definition. Many public agencies have statutory liability for a dangerous condition of public property. A dangerous condition is a condition of public property that creates a substantial risk of harm to persons using due care. To impose liability, a public agency must have notice of the dangerous condition for a sufficient length of time in order to have either remedied or warned of the condition. Using the term “dangerous” or “danger” in almost any context in the HSM2 would give rise to serious exposure to tort liability and is to be avoided. An objective and neutral description of the condition lessens that exposure.
• Defect/defective. When these terms are used to describe a feature or condition of a roadway (e.g., “roadway defect,” “defective condition”), the writer has essentially concluded that a dangerous condition exists for which an agency may be liable. These terms have the added legal implication of product liability, which is to say liability without fault. A maker of a product (e.g., a Ford Pinto) is liable without fault if that product is defective and causes damage. In other words, a plaintiff does not have to prove negligence in the making of the product, only that it was defective. A plaintiff would argue that a transportation agency that has a roadway defect or a defective condition of a roadway is liable for any injuries to a roadway user caused by that defect. While a transportation agency has not yet been considered a maker of a product within the concept of product liability, using the terms “defect” or “defective” gratuitously connotes legal implications that are to be avoided.
• Drop off. This is a term used to describe a condition at the edge of pavement where there is a difference in elevation from the pavement to the shoulder. The term “drop off” however does not have a benign connotation. It implies a hazard that would cause a driver who encounters it to lose control of a vehicle. The neutral phrase “elevation difference between the pavement and the shoulder” describes what the condition is. The amount of that elevation difference, if large enough and under certain circumstances, may pose a challenge for a driver to negotiate the vehicle back onto the traveled way. The elevation differential can be measured, recorded, and utilized without the use of the subjective and pejorative term “drop off.”
• Hazard/hazardous. These terms are synonymous with “danger/dangerous.” While these terms are not as commonly used as legal words of art, it is well known they have the same effect. They raise immediate and serious tort liability implications for the same reasons that “danger/dangerous” do and should have no place in the HSM2. Yet the HSM’s roadside hazard rating is used to develop a CMF in Step 10 of the Predictive Method, notwithstanding that it is conceded to be subjective and does not conform to the type of quantitative measurements found elsewhere in the HSM. Measuring the effect of certain roadside features on the frequency and severity of crashes, without calling them “hazards,” will move these concepts away from those laden with tort liability implications to concepts that are neutral in tone and quantitative in nature. Such terms as “hazard” and “hazardous” should not be used in the HSM2.
• Hot spot/black spot. These are pejorative terms used to describe a location where there is a concentration of crashes. The terms conclude that there is something bad about the site before any analysis has begun. Describing sites as “hot” or “black” is gratuitously judgmental. The neutral term “crash concentration” does not bias the user and allows for an objective analysis. Such an analysis may result in a conclusion that the site may benefit from an improvement.
• Negative effect. Effects, impacts, or outcomes should not be characterized as negative or positive. If there is a need to express these effects, impacts, or outcomes in relative terms, they

• should be stated objectively (e.g., greater, larger, less than, smaller). Depending on the context, even the use of “higher” or “lower” can sometimes carry the judgment of “better” or “worse.”
• Obstacle/obstruction. These terms can apply to many features or anomalies in and around a roadway. They literally describe something that stands in the way. The terms “obstacle” and “obstruction” are pejorative descriptors of an object, device, or feature. They are also conclusionary in their implication of a threat to a driver. If steps are not taken to remove them or avoid them, there may be legal implications for an agency. Users are better served if these items are identified and located in neutral terms (e.g., “an object, device, or feature is located 2 ft. from the traveled way,” as opposed to “too close to the traveled way”).
• Problem. The term “problem” describes a matter or issue in a negative way. It is a subjective conclusion that a matter is having an adverse effect. A problem is something that needs to be solved; it cannot be ignored. A more objective approach would be to identify issues, matters, or conditions that may benefit from further analysis.
• Risk/risky/risk factors/safety risk/level of risk/higher-risk facilities. Each of these terms suffers from using a pejorative descriptor that is an element for establishing tort liability. Agency liability can be predicated upon a dangerous condition of public property, which is defined as a condition that creates a substantial risk of harm when used with due care. If one is measuring the relative risk created by certain roadway features in order to prioritize an improvement, this task is not unlike what a jury in a civil action does in determining whether a risk is substantial enough to constitute a dangerous condition. When an agency essentially assumes the jury’s fact-finding role in establishing a dangerous condition, it makes a plaintiff’s burden to prove liability much easier. On the other hand, it is known or can be calculated that certain roadway features or characteristics have associated with them a measurable potential for a crash. These data can be used with other factors that may be considered to calculate the safety performance of a roadway segment. The word “risk” adds an unnecessary element that introduces tort liability implications.
• Trap. This is a gratuitously pejorative descriptor of a condition of a roadway facility. It connotes a concealed danger—one that would not be reasonably apparent to, and would not have been anticipated by, a person exercising due care. Because of its serious liability implications, the word “trap” does not belong in the HSM2.
• Unsafe. This word was discussed previously as one that expresses a personal opinion rather than one supported by data. It is also ambiguous in that it can refer to personal safety as well as the safety performance of a roadway. However, when used in the instant context, it is shorthand for describing a dangerous condition. It is an ultimate legal conclusion that bears no relevance to a data-driven engineering analysis, and it is to be avoided.
• Unshielded/unprotected. These descriptors are usually associated with obstacles and obstructions (discussed previously). If a roadway feature is described and located, it will be clear from its description whether it is shielded or protected. To describe it as “unshielded” or “unprotected” implies that there is a duty to do so and that duty has not been met. It unnecessarily raises a liability issue of failure to meet a standard of care. Users are better served with a neutral and factual description of the object. After an analysis, the object may be a candidate for a shielding strategy.
• Weak/weakness. The terms “weak” and “weakness” describe the relative strength of something in a pejorative way. (e.g., “While this methodology does not require a significant amount of data to apply it, its ‘weakness’ is in the reliability of results.”) It implies that there is something inherently wrong. It is not neutral and denigrates that to which it refers. A more neutral term would be “limitation” or, in some cases, “limited.” (e.g., “Below is a list of the strengths and limitations of each methodology.”)

Examples

It is useful to look at some examples of the language changes made during the review process of the HSM, both for the identification of terms and phrases to be avoided and for how they were revised accordingly.

An early draft of the HSM set forth the following in Chapter 4, Section 4.1.1, Page 11-3, line numbers 21–28:

21 Roads that intersect at angles less than 90° can lead to sight distance and

22 operational concerns. . . . Vehicles have a longer distance to travel through the

23 intersection thus increasing their exposure to conflicts, crossing distances for

24 pedestrians are increased and drivers, particularly older drivers, may find it difficult

25 to turn their head and neck far enough to view approaching vehicles. Vehicles

26 turning right at an acute angle may have difficulty keeping within the designated

27 lane. When right-turn-on-red (RTOR) are permitted, drivers may have more

28 difficulty judging gaps when turning.

The reviewer addressed the issues raised by the original text by noting the use of terms of personal opinion (“difficulty”/“difficult”) and those that are pejorative or prejudge an issue without supporting data (“concern”). Moreover, rather than simply stating in a positive way what the safety impact of reducing skew angle may be attributed to, it states negatively that skew can cause “sight distance and operational concerns.” It then proceeds to give examples of those “concerns,” which give rise to tort liability issues. Instead, the discussion should focus on what reducing skew angle may achieve.

The HSM went through several revisions before it was published. The foregoing material ended up in Chapter 14.6: Crash Effects of Intersection Design Elements, Section 14.6.2.1, Reduce Intersection Skew Angle, Page 14–16 as follows:

An intersection that is closer to perpendicular reduces the extent to which drivers must turn their head and neck to view approaching vehicles. Reducing the intersection skew angle can be particularly beneficial to older drivers, and can also result in increased sight distance for all drivers. Drivers may then be better able to stay within the designated lane and better able to judge gaps in the crossing traffic flow. . . . Reducing the intersection skew angle can reduce crossing distances for pedestrians and vehicles, which reduces exposure to conflicts.

The final text is a marked improvement from the original. It covers all the issues raised in the original text but restates them in a positive way by setting forth the benefits of reducing skew angle.

An early draft of the HSM2 provided another example of how to revise text containing terms and phrases to be avoided. Chapter 4: Network Screening, Section 4.2.3, Step 3 stated the following:

When using a systemic approach to network screening, this step involves identifying risk factors, which are traffic characteristics or roadway geometric characteristics that are associated with the focus crash types and/or severities. Risk factors are not necessarily the cause of crashes, but are simply associated with the occurrence of crashes. For example, if roadway departure crashes are the focus crash type, shoulder width less than four feet may be identified as a risk factor for those crashes. The systemic network screening process would then screen the network based on the presence of roadway shoulders less than four feet in width to identify locations with higher risk for roadway departure crashes.

The following sections describe the process for selecting performance measures for a site-based screening approach or risk-factors for a systemic-based screening approach.

The reviewer commented on the use of the terms “risk” and “risk factors.” These are not neutral terms. Rather, they are pejorative and unnecessarily interject the tort liability concept of risk into what should be an otherwise straightforward description of the traffic characteristics or roadway geometric characteristics associated with certain crash types or severities. A revision eliminating the word “risk” from the text addressed the comment.

When using a systemic approach to network screening, this step involves identifying traffic characteristics or geometric characteristics that are associated with the focus crash types and/or severities. These traffic characteristics or geometric features are not necessarily causes of crashes but are simply

associated with the occurrence of crashes. For example, if roadway departure crashes are the focus crash type, a shoulder width less than four feet may be identified as potentially associated with the occurrence of those crashes. This means a shoulder width of less than four feet is a common characteristic at the locations where the roadway departure crashes occur, but close review of crash reports may indicate that a primary contributing factor to the crashes was some other factor, such as driver fatigue. The systemic network screening process would then screen the network based on the presence of roadway shoulders less than four feet in width to identify locations with a higher potential for roadway departure crashes.

The following sections describe the process for selecting performance measures for a site-based screening approach or identifying common characteristics for a systemic-based screening approach.

The revised text clearly sets forth the concept of a systemic approach to network screening without creating tort liability issues.

Another example from an early draft of the HSM2 was found in Chapter 6: Select Countermeasures, Section 6.2.2, Contributing Factors for Consideration, as set forth on the following:

Possible contributing factors for the following types of crashes along roadway segments include: . . .

Fixed object

• Obstruction in or near roadway
• Inadequate lighting
• Inadequate pavement markings
• Inadequate signs, delineators, guardrail
• Slippery pavement
• Roadside design (e.g., inadequate clear distance)
• Inadequate roadway geometry
• Excessive speed

The reviewer commented that most of these possible contributing factors are preceded by an unsupported opinion in the form of a pejorative adjective that acts as a limitation on the scope of the factor. For example, “inadequate” would appear to preclude “confusing” (e.g., pavement markings) or “too many” (e.g., signs). Deleting the adjectives or using neutral descriptors, as set forth in the following, presents the user with a broader, objective set of factors to consider.

Possible contributing factors for the following types of crashes along roadway segments include: . . .

Fixed object

• Object in or near roadway
• Lighting
• Pavement markings
• Signs, delineators, guardrail
• Pavement surface
• Roadside design (e.g., clear distance)
• Roadway geometry
• Speed

These revised possible contributing factors now include many different types of situations where the factors may come into play. Moreover, they do not predispose a user into only looking at one aspect (e.g., “slippery” pavement as opposed to “crumbling,” “uneven,” or “potholes”).

The second draft of Chapter 2: Human Factors in the HSM2 also provides examples of terms and phrases to be avoided.

Users with normal vision, but poor contrast sensitivity, may have to get very close to a low-contrast target before detecting it.

Driver alertness on long tangent sections is often a concern, especially in rural areas.

Perception-reaction time—The amount of time and distance needed by one user to respond to a stimulus (e.g., hazard in road, traffic control device, or guide sign) depends on human elements, including information processing, driver alertness, driver expectations, and vision.

Reviewer comments pointed out that “poor” is a vague and judgmental term and that a more objective term should be used. “Often” is equally vague and frequency does not even appear to be

A CMF representing the impact on the frequency and severity of crashes of the entire influence area would also be categorized as. . . .

As is evident from these revisions, use of the term “safety” more sparingly usually results in clearer text. These are but a few examples of how the terms and phrases in this style and usage guide can be used to improve the content of the HSM2.

Conclusion

The research team kept these terms and phrases in mind when drafting and revising content in the HSM2. Of particular significance will be the context in which these terms and phrases appear. Context can change an otherwise innocuous word into one that may pose serious legal ramifications. This list is by no means exhaustive, and it is expected that more terms and phrases may be added to this style and usage guide.