Safety Impacts of Intersection Sight Distance (2018)

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.

NCHRP 17-59 59 CHAPTER 3: FINDINGS AND APPLICATION OVERVIEW Chapter 3 presents the model estimation results that describe the relationship between target crash frequency, available ISD, and other intersection characteristics. Results are presented for target crashes and target fatal and injury crashes (i.e., a subset of target crashes that involve at least one fatality or injury of any level). The safety effects of available ISD differ by traffic volume and speed limit on the major road; therefore, practitioners should use the results associated with a specific major road AADT and a specific major road speed limit when possible. If major road AADT and speed limit are not available, then practitioners may use the results presented for average conditions. The remainder of this chapter is divided into three sections. The first section presents the regression model estimation results and an interpretation of the findings. The second section translates these results into practical crash modification functions (CMFunctions), including step-by-step instructions and examples to support the application of results. The final section describes the process used to develop a practice-ready guidance document, which included thorough vetting by state, local, and regional practitioners with experience in the collection and analysis of ISD. The end result is a guidance document for practitioners to support decision- making about ISD. FINAL MODELS This section presents the final models for target crashes and target fatal and injury crashes. Models for target fatal and incapacitating injury crashes, target angle crashes, and target daytime crashes were also explored and uncovered similar trends between the frequency of these crash types and ISD. In some cases, the statistical significance of estimated model parameters decreased due to the smaller numbers of these more refined target crash type definitions. Target Crashes Table 17 presents the final model for target crashes considering available ISD. The following provides definitions of the three main variables capturing the safety effects of available ISD and interpretations of the estimated parameters. Speed Limit/ISD interaction (spdlmt/ISD): This variable represents an interaction between major road speed limit and available ISD. It is the key model variable capturing the safety effects of available ISD. Its value is the major road speed limit divided by the available ISD for a subject minor road/major road approach combination. The functional form of the variable, combined with the positive estimated model parameter, indicate the following:  The expected number of target crashes increases as available ISD decreases.  ISD is associated with expected target crash frequency in a non-linear fashion. The sensitivity of the expected number of target crashes to changes in ISD is highest when

NCHRP 17-59 60 ISD is shorter, and decreases as ISD increases (i.e., the safety benefit of increasing ISD from 300 to 600 feet is substantially larger than the safety benefit of increasing ISD from 1000 to 1300 feet).  The sensitivity of the expected number of target crashes to changes in ISD increases as the speed limit on the major road increases (e.g., the safety benefit of increasing ISD from 300 to 600 feet is expected to be substantially larger on high speed roads compared to low speed roads). Major road AADT/ISD interaction (LmajAADT/ISD): This variable represents an interaction between major road two-way AADT and ISD. Its value is zero if the major road AADT is greater than 5,000. Its value is one divided by available ISD for the applicable minor road approach direction (i.e., the inverse of available ISD) if the major road AADT is less than or equal to 5,000. The negative parameter represents the finding that the effect of ISD on the expected number of target crashes is smaller on lower volume roads than on higher volume roads. Major road AADT/ISD interaction (MmajAADT/ISD): This variable represents a second interaction between major road two-way AADT and ISD. Its value is zero if the major road AADT is less than or equal to 5,000 or greater than 15,000. Its value is one divided by available ISD for the applicable minor road approach direction (i.e., the inverse of available ISD) if the major road AADT is greater than 5,000 and less or equal to 15,000. The negative parameter represents the finding that the effect of ISD on the expected number of target crashes is smaller on medium-volume roads than on higher volume roads. However, the parameter is less negative than for the low volume interaction, representing the finding that the effect of ISD on the expected number of target crashes is larger on medium-volume roads than on lower volume roads. The following provides definitions and interpretations of the estimated parameters for other variables in the model. log-majAADT: The natural logarithm of the major road AADT has a positive estimated parameter of 0.244. The expected number of target crashes increases as major road AADT increases. log-minAADT: The natural logarithm of the minor road AADT has a positive estimated parameter of 0.536. The expected number of target crashes increases as minor road AADT increases. The estimated main effect of minor road traffic volume is stronger than the main effect of major road traffic volume, likely due to the nature of the target crash types, as well additional major road volume effects captured through a major road AADT and ISD interaction. Left-Direction Indicator (LT): With the analysis unit being an approach direction, this variable takes a value of one if the direction being considered is to the left of the minor approach (i.e., approaching vehicles are from the left) and zero otherwise. The positive parameter indicates that the left direction from the approach is associated with higher target crash frequency than the right direction.

NCHRP 17-59 61 Table 17. Model Estimation Results for Expected Number of Target Crashes. Variable Coefficient Standard Error Z-Score P-Value Constant -8.147 1.382 -5.90 <0.001 β1 (log-majAADT) 0.244 0.120 2.04 0.042 β2 (log-minAADT) 0.536 0.072 7.42 <0.001 β3 (LmajAADT/avaiISD) -243.0 185.3 -1.31 0.190 β4 (MmajAADT/avaiISD) -177.8 86.53 -2.06 0.040 β5 (LT) 0.334 0.120 2.79 0.005 β6 (fourleg) 0.845 0.126 6.68 <0.001 β7 (median) -0.016 0.166 -0.09 0.925 β8 (spdlmt) -0.021 0.008 -2.75 0.006 β9 (spdlmt/ISD) 7.194 2.450 2.94 0.003 β10 (grd500) -0.061 0.026 -2.32 0.021 β11 (lnT) 1 (exposure) Dispersion (α) 2.66 0.278 N = 1653; Log-likelihood = -1296.89; Pseudo R2 = 0.055 Four-leg Intersection Indicator (fourleg): This variable takes a value of one if the intersection is four-leg and zero if three-leg. The positive parameter indicates that four-leg intersections are associated with higher target crash frequency than three-leg intersections. Median: This variable takes a value of one if there is a median present on the major road and zero otherwise. Median presence was expected to decrease the target crash frequency since, depending on median width, it would allow left-turning drivers to take refuge in the median (i.e., cross one roadway at a time). While this effect did exist, it was insignificant. The model parameter was still reported to demonstrate this result. Grade on the major approach at 500 feet from the intersection (grd500): This variable represents the measured vertical grade on the major road approach 500 feet prior to the intersection. The model presented in Table 17 was used to create a CMFunction, which can be used to calculate the effect of changing available ISD for an approach direction. The CMF calculation for total target crashes uses this CMFunction and is presented in Equation 3. ܥܯܨ் ൌ ஼ெி೅భ஼ெி೅మ (3)

NCHRP 17-59 62 where: ܥܯܨ் = CMF for target crashes for an approach direction. ܥܯܨ்೔ = Target crash CMF for condition of interest i (where i = 1 for proposed condition and i = 2 for existing condition). Equation 4 provides the CMFunction used to determine the target crash CMF for the proposed or existing condition. ܥܯܨ்೔ ൌ ୣ୶୮ ሺି଴.଴ଶଵൈ௉ௌ௅ାళ.భవరൈುೄಽ಺ೄವ೔ ା షమరయ.బబవൈಽ೚ೢಲಲವ೅೘ೌೕ ಺ೄವ೔ ା షభళళ.ఴమలൈಾ೔೏ಲಲವ೅೘ೌೕ ಺ೄವ೔ ሻ ୣ୶୮ ሺି଴.଴ଶଵൈ௉ௌ௅ାళ.భవరൈುೄಽ಺ೄವ್ೌೞ೐ ା షమరయ.బబవൈಽ೚ೢಲಲವ೅೘ೌೕ ಺ೄವ್ೌೞ೐ ା షభళళ.ఴమలൈಾ೔೏ಲಲವ೅೘ೌೕ ಺ೄವ್ೌೞ೐ ሻ (4) where: PSL = Posted speed (in mph). LowAADTmaj = 1 if major road AADT ≤ 5,000; otherwise 0. MidAADTmaj = 1 if 5,000 < major road AADT ≤ 15,000; otherwise 0. ISDi = Proposed or existing available intersection sight distance for the condition of interest i (where i = 1 for proposed condition and i = 2 for existing condition) (in feet). ISDbase = Base intersection sight distance for an approach direction (in feet). For practical applications, this value is assumed to be 1,320 feet. Target Fatal and Injury Crashes Table 18 presents the final model for target fatal and injury crashes considering available ISD. The general trends and directions of safety effects were consistent with the target crash model reported in Table 17 with the following exceptions:  The smaller sample sizes corresponding to target fatal and injury crashes did not allow the fatal and injury model to distinguish between “low” and “medium” volume regions in the major road volume/ISD interaction. Therefore, a different interaction was created that combines low and medium volumes (LMmajAADT/avaiISD). Its value is zero if the major road AADT is greater than 15,000. Its value is one divided by available ISD for the minor road major road approach combination (i.e., the inverse of available ISD) if the major road AADT is less than or equal to 55,000. The negative parameter still represents the finding that the effect of ISD on the expected number of target fatal and injury crashes is smaller on lower- and medium-volume roads than on higher volume roads.  The sign of the median variable, included to capture possible safety effects resulting from left-turning drivers being able to take refuge in the median, changed but the estimated parameter remained statistically insignificant.

NCHRP 17-59 63 The other key ISD-related findings of the target crash model remain the same for the target fatal and injury crash model, including the following:  The expected number of fatal and injury target crashes increases as available ISD decreases.  ISD is associated with expected target fatal and injury crash frequency in a non-linear fashion. The sensitivity of the expected number of target fatal and injury crashes to changes in ISD is highest when ISD is shorter, and decreases as ISD increases.  The sensitivity of the expected number of target fatal and injury crashes to changes in ISD increases as the speed limit on the major road increases. Interpretations of estimated model parameters corresponding to the main effects of major and minor road traffic volumes, left turn approach direction, number of intersecting legs, speed limit, and grade on the major road approaches are consistent with those from the target crash model. Table 18. Model Estimation Results for Expected Number of Target Fatal and Injury Crashes. Variable Coefficient Standard Error Z-Score P-Value Constant -8.234 1.431 -5.75 <0.001 β1 (log-majAADT) 0.115 0.117 0.98 0.325 β2 (log-minAADT) 0.498 0.090 5.55 <0.001 β3 (LMmajAADT/avaiISD) -155.5 108.6 -1.43 0.152 β4 (LT) 0.507 0.152 3.33 0.001 β5 (fourleg) 0.953 0.162 5.89 <0.001 β6 (median) 0.215 0.210 1.03 0.305 β7 (spdlmt) -0.009 0.010 -0.95 0.340 β8 (spdlmt/ISD) 6.335 2.915 2.17 0.030 β9 (grd500) -0.054 0.035 -1.54 0.125 β10 (lnT) 1 (exposure) Dispersion (α) 3.136 0.155 N = 1653; Log-likelihood = -820.79; Pseudo R2 = 0.049 The model presented in Table 18 was used to create a CMFunction for target fatal and injury crashes, which can be used to calculate the effect of changing available ISD for an approach direction. The CMFunction for target fatal and injury crashes uses this CMFunction and is presented in Equation 5.

NCHRP 17-59 64 ܥܯܨ்ிூ ൌ ஼ெி೅ಷ಺భ஼ெி೅ಷ಺మ (5) ܥܯܨ்ிூ = CMF for target fatal and injury crashes for an approach direction. ܥܯܨ்ிூ೔ = Target fatal and injury crash CMF for condition of interest i (where i = 1 for proposed condition and i = 2 for existing condition). Equation 6 provides the CMFunction used to determine the target fatal and injury crash CMF for the proposed or existing condition. ܥܯܨ்ிூ೔ ൌ ୣ୶୮ ሺି଴.଴଴ଽൈ௉ௌ௅ାల.యయఱൈುೄಽ಺ೄವ೔ ା షభఱఱ.ఱబరൈಽ೚ೢಾ೔೏ಲಲವ೅೘ೌೕ ಺ೄವ೔ ሻ ୣ୶୮ ሺି଴.଴଴ଽൈ௉ௌ௅ାల.యయఱൈುೄಽ಺ೄವ್ೌೞ೐ ା షభఱఱ.ఱబరൈಽ೚ೢಾ೔೏ಲಲವ೅೘ೌೕ ಺ೄವ್ೌೞ೐ ሻ (6) where: LowMidAADTmaj = 1 if major road AADT ≤ 15,000; otherwise 0. Table 19 provides descriptive statistics for the variables as they were specified in the final target and target fatal and injury crash models. Descriptive statistics for the actual variables (without transformations or interactions) were provided in Table 11 through Table 16. Table 19. Descriptive Statistics for Variables in Target Crash Models. Variable Mean Standard Deviation Minimum Maximum log-majAADT 9.0046 0.9350 5.6058 10.482 log-minAADT 6.7550 0.9425 3.7317 9.6574 LmajAADT/avaiISD 0.0003 0.0006 0 0.0101 MmajAADT/avaiISD 0.0006 0.0010 0 0.0100 LMmajAADT/avaiISD 0.0009 0.0010 0 0.0101 LT 0.5021 0.5001 0 1 fourleg 0.5214 0.4997 0 1 median 0.2807 0.4495 0 1 spdlmt 48.206 9.6412 20 70 spdlmt/ISD 0.0571 0.0324 0.1893 0.3535 grd500 -0.93 2.2 -8.7 7.9 The previously defined ISD quality variables were tested in the model specifications, but ultimately not included in the final model specifications. Trends indicated that the expected numbers of target and target fatal and injury crashes was lowest when ISD quality was rated as

NCHRP 17-59 65 “3.” This finding supports the hypothesis that objects in the roadside could be beneficial in helping drivers that are stopped on the minor road estimate the speeds of and distances to approaching vehicles on the major road approaches. Future research to build on this finding is needed. SUMMARY In summary, the expected number of target crashes is associated with available ISD. Target crash frequencies increase as available ISD decreases. Results of this research also suggest that ISD is associated with expected crash frequency in a non-linear fashion. The sensitivity of the expected number of target crashes to changes in ISD is highest when ISD is shorter, and decreases as ISD increases (i.e., the safety benefit of increasing ISD from 300 to 600 feet is substantially larger than the safety benefit of increasing ISD from 1,000 to 1,300 feet). The results also suggest that the impacts of ISD on crash frequencies vary as a function of the major road traffic volume and the major road speed limit. The sensitivity of the expected number of crashes to changes in ISD increases as traffic volume and speed limit increase. CMFunctions for each of the target crash types were estimated using the regression models. Applicability of the CMFunctions varies by available ISD, major road AADT, and speed limits available in the dataset. Operational and geometric characteristics found to be significantly associated with higher crash frequency (i.e., a unit increase in the variable is associated with an increase in crash frequency) include the following:  Major road AADT.  Minor road AADT.  Four-leg intersection indicator (compared to three-leg intersection).  Left-direction indicator (compared to right direction).  Vertical grade on the major road approach.  Speed limit on the major road approach. The presence of a median was expected to provide a safety benefit for the target crashes since, depending on median width, its presence would allow left-turning drivers to take refuge in the median (i.e., cross one roadway at a time). However, the modeling effort was not able to uncover this relationship. The CMFunctions presented in the Final Model section are relatively complex and depend on available ISD, major road AADT, and posted speed. In addition to presenting the CMFunctions, the project team, with direction from the project panel, translated the results into twelve charts for practitioners in design, planning, operations, and traffic safety concerned with improving intersection safety. The charts provide safety guidance for three separate bins of major road AADT for target crashes and two separate bins of major road AADT for target fatal and injury crashes. The project team identified the following three major road AADT ranges for target crashes:

NCHRP 17-59 66  Low major road AADT. The major road AADT is less than or equal to 5,000.  Mid major road AADT. The major road AADT is greater than 5,000 but less than or equal to 15,000.  High major road AADT. The major road AADT is greater than 15,000. The project team divided the major road AADT into two categories for target fatal and injury crashes. The low-mid major road AADT category consists of major road AADT less than or equal to 15,000. The high major road AADT category consists of major road AADT greater than 15,000. Charts for each crash type are separated by posted speed limit, ranging from 35 to 60 mph. This yields 6 charts for each crash type, or 12 total charts. Caution should be used for posted speeds outside the 35 to 60 mph range. Practitioners can apply these charts to estimate the following:  The expected change in crash frequency at intersection approach directions (i.e., left direction or right direction), approach level (i.e., for one minor road), or intersection level (i.e., for the intersection as a whole) from changing the available ISD.  The safety effects of changing available ISD at existing intersections.  The safety effects of changing available ISD during the planning and design stages of the project development process. The following are several notes of importance regarding the charts presented in this section. 1. Users will select a chart based on the crash type of interest (target or target fatal and injury crashes) and the major road posted speed (e.g., 55 mph). CMFunctions may be applied directly, but caution should be used when considering the range of posted speed and available ISD. Note that the base ISD condition is 1,320 feet for the charts. 2. The output from the charts is a CMF that applies to a specific minor road approach direction, which requires the analyst to know the initial travel directions of vehicles involved in crashes at the study intersection. 3. Approach direction CMFs can be combined using a weighted average based on the number of crashers per approach direction to develop an intersection level CMF. An example in the guidance document presents the methodology for combining CMFs to develop an intersection level CMF for target crashes. (See NCHRP Research Report 875: Guidance for Evaluating the Safety Impacts of Intersection Sight Distance.) Additionally, Chapter III of the guidance document and one of the examples may be used to estimate an intersection level CMF for total crashes. 4. The CMFunctions and charts treat major road AADT as grouped data. More precise AADTs cannot be used to further refine CMF estimates. 5. The maximum available ISD that may be used is 1,320 feet (i.e., a quarter mile). The data collection were truncated at this point as the project team and panel felt no additional

NCHRP 17-59 67 benefit may be realized for ISD beyond this value. The value 1,320 is used as the base condition for the charts, but direct comparisons may be made between two alternative CMFs on the chart. Figure 10 includes a sample chart to illustrate the step-by-step instructions with corresponding numbers that align with the following steps:  Step 1. Identify the minor road approach sight distance of interest (left-looking or right- looking) for which a new ISD is proposed. Measure the sight distance along the major road for that direction using the method in Chapter 2 (of the guidance document). This is the existing ISD for the approach and direction.  Step 2. Identify the type of crash of interest for the analysis: Target or Target Fatal or Injury Crash. o Target crashes are defined as a crash involving a vehicle from the major road and a vehicle entering from the minor road. o Target fatal and injury crashes are a subset of target crashes that involve one or more injuries or fatalities.  Step 3. Identify the appropriate chart type based on the type of crash and major road posted speed.  Step 4. Identify the curve that corresponds with the appropriate major road AADT range. o Three ranges of major road AADT are used for target crashes. o Two ranges of major road AADT are used for target fatal and injury crashes.  Step 5. Using the selected curve, plot the existing and the proposed ISD. If any ISD exceeds 1,320 ft, use 1,320 ft in the chart as this is the maximum.  Step 6. Calculate the CMF for changing the site distance from the existing to the proposed by dividing the CMF from the chart for the proposed ISD by the CMF from the chart for the existing ISD.  Step 7. Apply the resulting CMF to the crashes of interest associated with that direction.

NCHRP Figure 11 findings crashes a and injur example the guida Guidance 17-59 through Fi for target an nd differ by y crashes an application nce docume for Evalua gure 22 prov d target fata major road d also differ for a one dir nt develope ting the Safe Figure 10 ide the fina l and injury posted spee by major ro ection sight d as part of ty Impacts o 68 . Sample C l charts deve crashes. Fig d. Figure 17 ad posted s distance up this research f Intersecti hart. loped for im ure 11 throu through Fig peed. The n grade. Furth (See NCHR on Sight Dis plementati gh Figure 1 ure 22 are f ext section p er examples P Research tance.) on of the saf 6 are for tar or target fat rovides an are provide Report 875 ety get al d in :

N CHRP 17-59 Figure 11. CMF for Target Cra 69 shes when Posted Speed Equ als 35 mph.

N CHRP 17-59 Figure 12. CMF for Target Cra 70 shes when Posted Speed Equ als 40 mph.

N CHRP 17-59 Figure 13. CMF for Target Cra 71 shes when Posted Speed Equ als 45 mph.

N CHRP 17-59 Figure 14. CMF for Target Cra 72 shes when Posted Speed Equ als 50 mph.

N CHRP 17-59 Figure 15. CMF for Target Cra 73 shes when Posted Speed Equ als 55 mph.

N CHRP 17-59 Figure 16. CMF for Target Cra 74 shes when Posted Speed Equ als 60 mph.

N CHRP 17-59 Figure 17. CMF for Target Fatal and In 75 jury Crashes when Posted Sp eed Equals 35 mph.

N CHRP 17-59 Figure 18. CMF for Target Fatal and In 76 jury Crashes when Posted Sp eed Equals 40 mph.

N CHRP 17-59 Figure 19. CMF for Target Fatal and In 77 jury Crashes when Posted Sp eed Equals 45 mph.

N CHRP 17-59 Figure 20. CMF for Target Fatal and In 78 jury Crashes when Posted Sp eed Equals 50 mph.

N CHRP 17-59 Figure 21. CMF for Target Fatal and In 79 jury Crashes when Posted Sp eed Equals 55 mph.

N CHRP 17-59 Figure 22. CMF for Target Fatal and In 80 jury Crashes when Posted Sp eed Equals 60 mph.

NCHRP Example A three-l 7,000 and change in condition Target C Using Fig ISD of 40 0.77. Thi vehicles Target C Use Equa 17-59 – One Dire eg intersecti a major ro average tar of 400 feet rashes – Gr ure 23, the 0 feet is 1.4 s factor app approaching rashes - CM tion 3 to co ction ISD U on with stop ad posted sp get crash fre to a propos aphical App CMF for the 6. Therefor lies to multi from the le Figu Function mpute the C pgrade control on eed of 55 m quency wh ed condition roach proposed I e, the target vehicle crash ft. re 23. Exam MF for targ 81 the minor ro ph. A practi en increasin of 750 feet SD of 750 f crashes CM es involvin ple 1 – Tar et crashes. ad approach tioner is int g left-lookin . eet is 1.13. T F for impro g vehicles f get crashes has a majo erested in es g ISD from he CMF fo ving ISD is rom the min . r road AAD timating the an existing r the existin 1.13/1.46, o or road and T of g r

NCHRP The follo  T  T  T  T  T The CMF The CMF The targe The targe involving CMF can Equation 17-59 wing condit he posted sp his site uses he proposed he existing I he base ISD for the pro for the exi t crash CMF t crashes CM vehicles fr be compute 10. ions apply: eed is 55 m the MidAA ISD is 750 SD is 400 f is 1,320 fee posed ISD c sting ISD is is calculat F for impr om the mino d directly b ph. DT for the m feet. eet. t. ondition is s shown in Eq ed as shown (9) oving ISD i r road and v y using the 82 ajor road ( hown in Eq uation 8. in Equation s 0.77. This ehicles app existing ISD i.e., MidAA uation 7. 9. factor appli roaching fro as the base DTmaj = 1.0 es to multiv m the left. N ISD directl ). (7) (8 ehicle crash ote that the y, as shown (10) ) es in

NCHRP 17-59 83 GUIDANCE DOCUMENT DEVELOPMENT An additional intent of this project was to develop a practice-ready guidance document for application of the model results and any supplementary relationships identified as part of the crash analysis. The development of the guidance document considered the current practices of agencies, the intended audience, and the various situations where ISD is used (e.g., design, planning, operational changes, and safety remediation). The key components of the data and research results were condensed into a shorter, more reader- friendly guidance document. Specifically, the ISD data collection protocol developed for this research was incorporated as a recommended standardized method for practitioners obtaining ISD. Furthermore, the results of the modeling for target crashes and target fatal and injury crashes previously presented in this chapter were utilized for CMF applications. Examples were developed based on the research to demonstrate to practitioners how to implement the CMF graphs. To test the draft guidance document, the project team assembled a focus group of practitioners to test the usefulness, understandability, and ease of application of the results. This focus group included both state and local representatives. The project team presented the draft guidance document to the focus group via webinar, which allowed for interactive discussion and comment. Subsequently, the guidance document was refined and finalized. The following two sections provide a detailed discussion of the vetting process and development of final guidance document. Vetting Process The draft guidance document was reviewed by the project panel and then vetted by experts in the field. In order to accomplish the vetting process, 10 practitioners with field or management experience were identified to participate in a vetting seminar conducted via webinar. These individuals represented state, local, and regional professionals with experience in the collection and analysis of ISD. The project team solicited interest for participation in the webinar via email. Webinar participants were provided PDF versions of the draft guidance document and webinar slides prior to the meeting. The webinar was held on Tuesday, June 29, 2015 from 10:00 am until 11:30 am using Skype for Business, hosted by the project team. Of the 10 invitees, 8 participated in the conversation. Ms. Kim Eccles, Dr. Scott Himes, and Ms. Kara Peach of the project team hosted the webinar. Participants were provided with a brief overview of the project history and background, data collection and analysis, and a summary of each chapter of the guidance document. Then, the project team asked a series of targeted questions regarding the overall flow of the guidance document, specific information presented in each chapter, and other such questions. Finally, the project team facilitated an open discussion that allowed for additional comments and feedback. Overall, feedback from the participants was positive. They provided constructive and thoughtful comments that were considered by the project team afterward. Participants stated that they found

NCHRP 17-59 84 the guidance document to be well laid out and were pleased with the completeness and succinctness of the charts. Additionally, they reported that the guidance document would be helpful and applicable in the field. Although several invitees were not able to participate in the webinar, they provided comments via email. All comments were included in the review and update of the guidance document. Table 20 presents the names, organizations, roles, and level of participation in the guidance document review. Table 20. Vetting Webinar Participants. Name Organization Position Review Participation Daniel Armentrout Clinton County Road Commission Director of Engineering Emailed comments Danielle Deneau Oakland County Road Commission Traffic Engineer Webinar participant Michael McNeil Ohio Department of Transportation Division of Planning Webinar participant Nate Miller Upper Valley Lake Sunapee Regional Planning Commission Executive Director Webinar participant Keith Nichols Hampton Roads Transportation Planning Organization Senior Transportation Engineer Webinar participant Jed Niffenegger City of Raleigh Public Works Department Transportation Operations Director Webinar participant Carrie Simpson North Carolina Department of Transportation Traffic Safety Project Engineer Webinar participant; emailed comments Derek Troyer Ohio Department of Transportation Division of Planning Webinar participant The vetting webinar participants provided extensive feedback on the guidance document – both during the webinar and electronic correspondence after the webinar. While many of the comments were minor, there were several suggested revisions that required further consideration. Those incorporated into the final guidance document include the following descriptions. Guidance pertaining to left-turn crashes was removed. Participants felt that if left-turn crashes are considered separately, then right turn crashes, angle crashes, etc., should be presented

NCHRP 17-59 85 separately. The project team agreed with participants that this specific crash type is too detailed for the purpose of the guidance document. With this in mind, results for left-turn crashes were removed and the focus of the document is target crashes and target fatal and injury crashes. The graphical representations of target crashes and target fatal and injury crashes were found to be confusing for some participants. One participant had a difficult time determining which graph to use for different circumstances. Other participants felt that practitioners in the field would not read the accompanying text and suggested more information be added to the graphs to explain their use. The project team decided not to overwhelm the graphics with additional details on their use, since practitioners in the field are not likely to be computing CMFs. The graphs are more likely to be used by practitioners in the office who have more background in the Highway Safety Manual methods. Since the guidance pertains to approach direction (i.e., for a specific ISD), participants felt that applying crashes to an approach direction ISD may be difficult for some practitioners. Webinar participants suggested that crash diagrams can be used to help with documenting crashes and applying them to an approach direction ISD consistently. The project team agreed and added an example crash diagram and an accompanying description in the crash history section. An example crash diagram and its application were also added to one of the example problems. Final Guidance Document The final guidance document is a result of several review stages, initiated with a draft outline submitted to the project panel for review in the second interim report. Feedback from the project panel was reviewed and the outline was further developed into a draft guidance document. The draft guidance document was once again reviewed by the project panel. The feedback received at that time was very helpful, and the suggestions were used to revise the draft guidance document. The vetting webinar and incorporating the suggested revisions provided at that time further refined the guidance document. Finally, a third review by the project panel was conducted and the guidance document was further refined based on their comments. The result of the condensed research summary, project panel comments, and vetting webinar revisions is a guidance document for practitioners to support decision-making about ISD. There are six chapters in the guidance document, as follows:  Chapter I (Introduction) provides an overview of the intended audience and uses, a brief summary of the guidance document history and background, instructions on how to use the guidance document, and important definitions and acronyms.  Chapter II (Measuring Sight Distance and Other Critical Information) outlines detailed recommended steps for measuring ISD and other important considerations, such as traffic volume and crash data.

NCHRP 17-59 86  Chapter III (Safety Performance and Intersection Sight Distance) provides practitioners information on the anticipated impact on crashes at an intersection from changing the sight distance. This chapter introduces a series of charts depicting CMFs for target crashes and target fatal and injury crashes. A sample chart and step-by-step instructions are also provided to help guide the practitioner.  Chapter IV (Examples) demonstrates how the charts can be used by the practitioner. Four sample problems are provided, each supported by a graphic illustration of the corresponding chart first introduced in Chapter III.  Chapter V (Other Countermeasures and Resources) briefly discusses additional considerations when examining ISD and resources for supporting guidance and instructions on ISD.  Chapter VI (Base Equations for Reference) provides the CMFunctions that resulted from the research and from which the charts in Chapter III are based. The guidance document is also supported by graphics to further assist the reader in their understanding. Light blue “Practitioner Tips” and dark blue “Tool Box” text boxes are used throughout the text to draw the readers’ attention to useful instructions or important definitions. The final guidance document is provided in NCHRP Research Report 875: Guidance for Evaluating the Safety Impacts of Intersection Sight Distance.