Application of Crash Modification Factors for Access Management, Volume 2: Research Overview (2021)

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.

40 Introduction The objectives of Task 3 were to (1) design and conduct a survey to identify transportation agency policies, practices, needs, and challenges related to quantifying the safety effects of access management features and (2) identify and review existing data sources to determine the avail- ability and quality of data to support this research. The survey results identified priority strategies. The data reconnaissance helped to determine the feasibility of evaluating priority strategies and strategy combinations. It also helped to identify potential data and methodological issues to con- sider during the data collection and analysis, thus informing the study design and data collection plan. This chapter describes the process and presents the results of the stakeholder survey and data reconnaissance efforts. Survey This section documents the methodology and results of a national survey conducted as part of this project. The survey included 10 questions to help identify the following: 1. The extent of quantification and tracking of safety effects of access management strategies, 2. The existence of policies or procedures for assessing the safety effects of access management strategies, 3. The level of analysis (i.e., site, intersection, or corridor) that would be most helpful, 4. Priority needs for research to estimate safety effects of individual or combination strategies, and 5. Priority access management strategies that should be the focus of the research. The survey was administered online using Survey Monkey for 2 weeks—October 2 to October 16, 2017. The survey initially targeted state department of transportation (DOT) traffic safety staff and was distributed to the AASHTO committees on safety, design, and traffic engineering. There were 30 survey respondents, representing 20 state DOTs. To expand the target audience to local agencies, the survey was redistributed for another 2 weeks—March 14 to March 31, 2018. The second round of the survey included the National Association of County Engineers (NACE); select metropolitan planning organizations (MPOs), i.e., MPOs with populations exceeding 1 million; and the ITE community. There were 19 additional survey respondents. Figure 13 shows the distribution and representation of respondents. The remainder of this section presents the tabulation of responses and survey findings, focusing on the initial survey of state agencies. Following the summary of initial survey responses is a com- parison and discussion of results from the redistributed survey. The survey results informed the gap analysis, study design, and guidance document. See Appendix A for the complete questions and Appendix B for responses to open questions. C H A P T E R 3 Data Reconnaissance and Assessment

Data Reconnaissance and Assessment 41   Summary of Initial Survey Responses Question 1: Do You Quantify the Safety Effects of Access Management Strategies To Support Related Decisions? Figure 14 shows that the majority of respondents do not quantify the safety eects of access management strategies to support the decision-making process. Of the survey respondents, 15 of 28 do not quantify the safety eects of access management strategies to support policy decisions. Comments to this question from respondents answering “Yes” include the following: • “Only for projects that request Safety Program funding (state or federal).” • “We quantify the eects of access management by using applicable CMFs from the clearing- house (national study’s)—we do not have any specic to Oregon.” • “We attempt, where possible, to use criteria from the Safety Manual, our Design Manual, and established access management elements to determine the access management criteria for projects and applications for access to our highways.” • “In the past, we have used an FHWA chart which shows the reduction in crashes based on number of access points along a corridor. is chart is now quite old, so an updated analysis would be helpful.” Comments to this question from respondents answering “No” include the following: • “. . . But we want to.” • “Only because we have not advanced to this level of analysis.” • “No, but access management aspects are included in the Development Coordination Manual and Road Design Manual.” STATE LOCAL OTHER NO RESPONSE Figure 13. Distribution and representation of survey respondents.

42 Application of Crash Modification Factors for Access Management Question 2: Do You Have a Policy or Procedure for Assessing the Safety Effects of Access Management Strategies? Figure 15 shows that 23 of 28 respondents to Question 2 indicated that their agencies do not have a policy or procedure for assessing the safety effects of access management strategies. This may indicate that even the agencies that quantify the safety effects of access management strategies may not have a standard procedure for doing so. Comments to this question from respondents answering “Yes” include the following: • “We probably do, related to the typical section choices . . . , but I don’t have the link.” • “Traffic safety does a good job detailing the benefits of RCIs [restricted crossing intersections]. I am not aware of any other procedure.” Yes (please describe in the Comments section below) No 0% 10% 20% 30% 40% 50% 60% Figure 14. Responses to Question 1: Do you quantify the safety effects of access management strategies to support related decisions? Yes (please provide a link or brief description in the comment box below) No 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% Figure 15. Responses to Question 2: Do you have a policy or procedure for assessing the safety effects of access management strategies?

Data Reconnaissance and Assessment 43   Comments to this question from respondents answering “No” include the following: • “We have general policy, but not specifically to the safety effects of access management.” • “Occasionally a researcher will gather data for some projects, but nothing that is routine.” Question 3: Do You Track Safety-Related Performance Measures After Implementing Access Management Strategies? Question 3 (see Figure 16) was intended to help identify whether there is an evaluation of safety after the implementation of access management strategies. Of the survey respondents shown in Figure 16, 21 of 28 indicated that they do not track safety metrics after implementing an access management technique. Comments to this question from respondents answering “Yes” include the following: • “Sometimes we conduct before-after analyses of batches of similar projects.” • “Only for projects that have been funded with HSIP [Highway Safety Improvement Program] federal funds as required by the yearly report. Currently, we provide a 3-year before and after analysis for the locations.” • “Although not for all applications, we have reviewed access management installations after the fact to determine effectiveness of the installation. We typically wait three years to make that determination.” • “Not as often as we should . . . Only if we do an after study.” • “We track all safety funded projects and report to the legislature.” • “On a limited basis, we have quantified the safety impacts related to access management efforts.” Comments to this question from respondents answering “No” include the following: • “No, we typically do not track safety-related performance measures after implementation; this isn’t part of our process. We mainly rely on determining the effectiveness of the strategies as identified in the ODOT’s CRF [Oregon DOT’s crash reduction factor] list.” • “We have done some studies but do not track anything official.” • “Only occasionally.” Yes (please describe in the Comments section below) No 0% 10% 20% 30% 40% 50% 60% 70% 80% Figure 16. Responses to Question 3: Do you track safety-related performance measures after implementing access management strategies?

44 Application of Crash Modification Factors for Access Management • “We have gone back and done some evaluations of access management projects, but we don’t track these projects statewide.” • “. . . Personally, I do not.” • “We may look at simple before and after for specific projects.” Question 4: What Is Your Priority Need to Estimate the Safety Effects of Access Management Strategies? Question 4 (see Figure 17) was intended to help identify whether the research priority should be individual access management strategies or combinations of strategies. Of the survey respon- dents shown in Figure 17, 15 of 27 indicated a priority need for combination strategies and 12 of 27 indicated a priority need for individual strategies. Some of the comments to this question reflected a difficulty in choosing which option should be prioritized. Some respondents expressed a desire to estimate the safety effects of both groups of access management strategies. Comments to this question from respondents answering “combination strategies” included the following: • “Would also be helpful, with public outreach, to have some background information stating how the effects of combination strategies can improve safety (for example, PowerPoint slides, handouts etc.)” • “Difficult selection—both might be necessary in any given situation. I could see where it might be difficult to attribute a percentage of the overall safety benefit to one strategy where multiple strategies are deployed as part of the same project.” • “With Access Management Plans we are identifying several safety improvements along/adjacent to a corridor. I think that showing the benefit of several improvements in concert will be more significant than showing the benefit of individual improvements.” Comments to this question from respondents answering “individual strategies” included the following: • “Given a finite timeframe and/or resources, individual strategies would be easier to evaluate.” • “Both are desirable but isolated improvements are more likely to be achievable.” Safety effects of individual strategies (e.g., relocate driveway, close median opening) Safety effects of combination strategies (e.g., install non-traversable median and consolidate driveways along a corridor) 0% 10% 20% 30% 40% 50% 60% Figure 17. Responses to Question 4: What is your priority need to estimate the safety effects of access management strategies?

Data Reconnaissance and Assessment 45   • “We attempt to use both of these elements, where applicable, equally.” • “At this point, the majority of our focus is on individual strategies, making it a higher priority, but combination strategies would be helpful as well.” • “Hard to pick one—both are site specific and are needed.” Question 5: What Are Your Priority Needs in Terms of the Level of Analysis? Question 5, shown in Figure 18, asked respondents to rank their priority needs from 1 to 3 (where 1 indicates the highest priority and 3 indicates the lowest) based on a spatial scale including the following levels: site level, such as an isolated median; intersection level, such as corner clear- ance criteria; and corridor level, such as converting undivided roadways to TWLTLs. Table 44 summarizes how respondents ranked each level of analysis. A majority of respondents indicated a priority of 1 for a corridor level of analysis. A majority of respondents indicated a priority of 2 for an intersection level of analysis. A majority of respon- dents indicated a priority of 3 for a site level of analysis. The corridor level of analysis appears to have the highest priority based on the survey responses. Question 6: What Are Your Priority Needs in Terms of Access Management Strategies? Table 45 summarizes how respondents ranked each access management strategy. Note that respondents were asked to rank each strategy individually with no limit on how many strategies could be categorized as high priority, medium priority, or low priority. Site (e.g., modify design of individual driveways or turn lanes, install isolated medians) Intersection (e.g., establish corner clearance criteria, install left-turn lanes, convert to roundabout, convert to restricted crossing U-turn) Corridor (e.g., manage location and spacing of access points, install medians, convert undivided to TWLTLs) 0 0.5 1 1.5 2 2.5 Score Figure 18. Responses to Question 5: What are your priority needs in terms of the level of analysis? How many respondents ranked each level as priority… 1 2 3 Site Level 5 9 15 Intersection Level 8 15 6 Corridor Level 16 5 8 Table 44. Survey summary: level of analysis.

46 Application of Crash Modification Factors for Access Management Access Management Strategy 1 High Priority 2 Medium Priority 3 Low Priority Total Weighted Average % responding and number of responses Alternative intersection and interchange design 55% 16 34% 10 10% 3 29 2.45 Control driveway design elements 38% 11 38% 11 24% 7 29 2.14 Control intersection design elements 17% 5 48% 14 34% 10 29 1.83 Convert two-way streets to one-way operation 7% 2 24% 7 69% 20 29 1.38 Establish corner clearance criteria 28% 8 59% 17 14% 4 29 2.14 Improve cross connectivity 28% 8 45% 13 28% 8 29 2 Install continuous TWLTL on undivided highway 24% 7 55% 16 21% 6 29 2.03 Install left-turn lanes 28% 8 52% 15 21% 6 29 2.07 Install non-traversable medians and accommodate left turns and U-turns 61% 17 32% 9 7% 2 28 2.54 Install right-turn lanes 14% 4 41% 12 45% 13 29 1.69 Install service or frontage roads 14% 4 48% 14 38% 11 29 1.76 Manage location and spacing of unsignalized access 59% 17 38% 11 3% 1 29 2.55 Manage spacing of traffic signals 28% 8 48% 14 24% 7 29 2.03 Manage location, spacing, and design of median openings and crossovers 62% 18 28% 8 10% 3 29 2.52 Manage the spacing of signalized and unsignalized access on crossroads in the vicinity of freeway interchanges 55% 16 34% 10 10% 3 29 2.45 Provide adequate sight distance at access points 21% 6 48% 14 31% 9 29 1.9 Table 45. Survey summary: priority of access management strategies. A majority of the respondents ranked the following five strategies as “high priority”: • Manage location, spacing, and design of median openings and crossovers; • Manage location and spacing of unsignalized access; • Install non-traversable medians and accommodate left turns and U-turns; • Alternative intersection and interchange design; and • Manage the spacing of signalized and unsignalized access on crossroads in the vicinity of freeway interchanges. A majority of the respondents ranked the following three strategies as “medium priority”: • Establish corner clearance criteria; • Install continuous TWLTL on undivided highway; and • Install left-turn lanes. One strategy received a ranking of “low priority” from 20 of the 29 respondents: “Convert two-way streets to one-way operation.” The strategy with the highest weighted average was “Manage location and spacing of unsignal- ized access.” It was followed by “Install non-traversable medians and accommodate left turns and

Data Reconnaissance and Assessment 47   U-turns,” “Manage location, spacing, and design of median openings and crossovers,” “Manage the spacing of signalized and unsignalized access in the vicinity of freeway interchanges,” and “Alternative intersection and interchange design.” The five strategies with the highest weighted average are also the strategies that were ranked as “high priority” by a majority of respondents. The two strategies that followed the top five in terms of weighted average are “Control drive- way design elements” and “Establish corner clearance criteria.” Question 7: Are There Other Access Management Strategies or Combination(s) of Strategies That Should Be Prioritized in This Project? Of the respondents to Question 7, 21 of 27 indicated “No.” Two additional strategies that were identified by respondents were • “Median installation, reduction of driveway density, and addressing open access within the influence area of an intersection.” (Note these are included in the list of potential strategies.) • “Avoiding the addition of a fourth leg at a three-legged intersection. Prohibiting movements at a signalized intersection so as to avoid additional phases.” Respondents also offered the following comments to this question: • “How do access management techniques impact pedestrians and bicyclists? Both from a crash perspective and from how people feel using these facilities?” • “Should the traffic volume of an access point affect the minimum spacing? For example, it seems that a high-volume access should be spaced farther from another high-volume access than from a low-volume access or between two low-volume accesses.” • “The effect of access management on non-motorized, e.g., increased exposure.” • “Partnering with local agencies where they might adopt and enforce the State’s Access Man- agement Guidance/Policies in order to provide effective access management at interchanges and intersections that involve state highways and local roads.” • “It would be good to quantify the general before and after benefits of displaced left turns as well as continuous flow T-intersections. I often wonder about the safety associated with the spacing of non-traditional intersections, e.g., since RCIs elongate an intersection, would the spacing needs be greater, or would they be less since the intersection operates more effectively? How would that impact safety? Aside from safety for cars, would the various strategies have positive or negative impacts on bike/ped traffic? Also, would bypass lanes generally show an increase or decrease in safety? For trucks needing to turn left from the minor leg to the major leg, do inside merging lanes work? There is definitely disagreement pertaining to this.” Question 8: May We Follow Up with You for Additional Information? If Yes, Please Provide the Following Information: Name, Organization, Email, Phone Table 46 lists the agency affiliations of respondents that replied “Yes” to Question 8. Question 9: Is There Someone Else in Your Organization You’d Like to Refer Us to for Additional Information? If Yes, Please Provide the Following Information: Name, Organization, Email, Phone Table 47 lists the organizations whose respondents replied “Yes” to Question 9. Question 10: Comments Question 10 was for open-ended comments. The following comments were provided by the respondents: • “Some of the treatments are already in the Highway Safety Manual (1st Edition) (i.e., left-turn lanes, right-turn lanes, TWLTL). These could be applied as different alternatives.”

48 Application of Crash Modification Factors for Access Management • “My use of this analysis is not very frequent.” • “Although ‘Install service or frontage roads’ is low for our agency mostly to avoid mainte- nance costs, there is an incident management and network connectivity benefit that appears to not be accounted.” • “This effort looks like it will be quite helpful as we continue to look at the benefit and costs of low cost/high benefit safety improvements. Thank You.” Summary of Responses to Second Survey The survey was redistributed to local agencies and practitioners, including NACE, MPOs, and the ITE community. The results of the second survey closely match the results of the initial survey. The following is a summary of the 19 additional responses. • Approximately 25 percent of the initial survey respondents and 31 percent of the second survey respondents track safety-related performance after implementing access management strategies. • 56 percent of the initial survey respondents and 53 percent of second survey respondents indicated that combination strategies have a slightly higher priority than individual strategies. • Both groups of survey respondents indicated that corridor strategies are the highest priority for research, followed by intersection strategies and site strategies. • The top five priority access management strategies were the same for both groups of survey respondents: – Manage location, spacing, and design of median openings and crossovers; – Manage location and spacing of unsignalized access; – Install non-traversable medians and accommodate left turns and U-turns; – Alternative intersection and interchange design; and – Manage the spacing of signalized and unsignalized access on crossroads in the vicinity of freeway interchanges. Agency North Carolina DOT Burns & McDonnell Iowa DOT Maine DOT Oklahoma DOT Arkansas DOT South Carolina DOT Oregon DOT New Hampshire DOT Kansas DOT North Dakota DOT Utah DOT Washington DOT Georgia DOT Mississippi DOT Illinois DOT Wisconsin DOT Minnesota DOT Minnesota DOT—Metro District Table 46. Survey summary: agency affiliations of respondents agreeing to provide additional information. Agency Oregon DOT Maine DOT Iowa DOT Arkansas DOT Utah DOT Mississippi DOT Wisconsin DOT Minnesota DOT Table 47. Survey summary: additional contacts.

Data Reconnaissance and Assessment 49   Conclusions from the Survey The survey results indicated the following regarding the quantification and tracking of safety effects of access management strategies and the existence of policy or procedures for assessing these effects: • A majority of the respondents do not quantify the safety effects of access management strate- gies to support the decision-making process. • More than 80 percent of the respondents indicated that their agencies do not have a policy or procedure for assessing the safety effects of access management strategies. This may indicate that even the agencies that quantify the safety effects of access management strategies may not have a standard procedure for doing so. • Approximately 75 percent of the respondents indicated their agencies do not track safety metrics after implementing an access management technique. Priorities that may be gleaned from the survey respondents and applied to this study are the following: • More than half of the respondents indicated a priority need for estimating the effects of com- bination strategies; less than half indicated a priority need for focusing on individual strate- gies. Some of the comments to Question 4 reflected a difficulty in prioritizing between the two choices. There were respondents who expressed a desire to estimate the safety effects of both groups of access management strategies. • Corridor-level analysis appears to have the highest priority based on the survey responses as opposed to segment/intersection- or site-level analysis. • A majority of the respondents ranked five access management strategies as “high priority.” These are also the access management strategies that have higher weighted averages: – Manage location, spacing, and design of median openings and crossovers; – Manage location and spacing of unsignalized access; – Install non-traversable medians and accommodate left turns and U-turns; – Alternative intersection and interchange design; and – Manage the spacing of signalized and unsignalized access on crossroads in the vicinity of freeway interchanges. Data Reconnaissance This section provides a summary of existing data sources, focusing on the availability and quality of data to support the research. Before asking what data are available, it is necessary to consider what data are required and desired to support the anticipated research. For rigorous crash-based evaluations, the Empirical Bayes before-after method is the current state of the practice to develop CMFs. This method would be challenging, however, as it would require implementation data (i.e., where and when the access management strategy was implemented), a reference group of similar unchanged sites, and crash and traffic volume data for all sites (both treatment and reference sites) before and after implementation for a large sample of locations. As noted in Chapter 2, this is one reason for the limited number of quality CMFs for select access manage- ment strategies. Since the Empirical Bayes before-after method was infeasible, at least for many of the strategies and gaps of interest, cross-sectional methods, such as multivariable regression models were required. To minimize new data collection efforts, the review focused on identifying existing datasets that could be repurposed and/or expanded for use in this effort. For example, an existing dataset of crash, traffic volume, and general roadway characteristics could be expanded upon through

50 Application of Crash Modification Factors for Access Management the collection of additional variables related to access management. Similarly, datasets with high- quality information on roadways and access management features (e.g., for a corridor-planning study) could be supplemented with crash data for further safety modeling. Potential datasets included those used by project team members for safety research either directly or indirectly related to access management, a review of completed safety research projects spon- sored by either NCHRP or FHWA, and a literature search using the TRID database. The focus was on datasets prepared in roughly the previous 5 years. This focus reflects a desire to use current data in future research, as well as difficulties in acquiring and/or adding additional variables to old data as circumstances change over time. In the review, answers to the following specific questions were sought: • What data are readily available in existing databases? Specific data of interest include crash, roadway, and traffic data. • How many miles of road and counts of intersections (by facility type) are available? • What is the availability of suitable reference locations? In some cases, access management strategies are applied to an entire corridor, and there may not be other similar corridors in the area that could serve as a reference group. • What is the quality of existing data? • Which agencies provided the data? • Who is the point of contact for acquiring the data? Table 48 provides a summary of findings from the data reconnaissance. The data sources in boldface were the priority for data collection in Task 7.

Data Source Level of Data (corridor, site, intersection) Variables in Dataset Miles or Locations by Facility Type Availability of Reference Sites Data Quality Agencies That Provided Data Notes DTFH61-09-C- 00026 Safety Evaluation of Access Management Policies and Techniques Corridor level. Data for individual median openings, driveways, and intersections are available with additional effort to query the data at that level. Data on spacing are available with additional effort to query the data at that level. Provides a mix of residential, commercial, and mixed-use corridors in suburban, urban, and urbanizing areas. Segment Data: # of lanes, median type, two-way vs one- way operation, speed limit, frontage type, presence of lighting, condition of pavement markings, visual clutter, internal cross connectivity, presence of frontage or backage road, # of driveways by type and permitted movements, # of unsignalized and signalized intersections by legs and permitted movements, # with right- and left-turn lanes on mainline, corner clearance for at least some signalized intersections, # of median openings with and without left- turn lane, interchange-related spacing. Crash Data: total, injury (KABC), turning, rear-end, and right-angle crashes. Up to 3 years of data available. Traffic Data: Major road annual average daily traffic (AADT) and commercial truck AADT. 600 total miles of collector and arterial corridors 718 median openings 3,478 unsignalized intersections 1,742 signalized intersections 14,780 driveways Sites were used for cross- sectional analysis. Depending on approach and topic, a subset may be relevant as reference sites. High– variables were field checked. North Carolina, Minnesota, California 2006–2008 A valuable database because data for extended corridors exist. Data are from 2006 to 2008, so it may be difficult to add additional data reliably. Table 48. Summary of data reconnaissance findings. (continued on next page)

Data Source Level of Data (corridor, site, intersection) Variables in Dataset Miles or Locations by Facility Type Availability of Reference Sites Data Quality Agencies That Provided Data Notes NCHRP Project 17-62 Site and intersection level. Corridor-level data may be created by aggregating adjacent sites together where possible. Rural Two-Lane Segments: lane width, shoulder width, shoulder type, roadside hazard rating, horizontal curvature, passing lanes, TWLTLs, grade, and AADT. Intersections: skew angle, left- turn lanes, right -turn lanes, lighting, and major and minor AADT. Rural Multi lane Segments: lane width, shoulder width, shoulder type, median width, lighting, and AADT Intersections: skew angle, left- turn lanes, right -turn lanes, lighting, major and minor AADT. Urban/Suburban Arterials Segments: median width, proportion with parking, fixed object density, number of driveways by type, lighting, speed limit, parking type, AADT. Intersections: left -turn lanes, right-turn lanes, lighting, no right-turn-on-red, major and minor AADT. Rural Two-Lane Segments 2U – 4,395 mi Rural Two-Lane Intersections 3ST – 755 4ST – 1,348 4SG – 63 Rural Multi-lane Segments 4U – 645 mi 4D – 665 mi Rural Multi-lane Intersections 3ST – 562 4ST – 570 4SG – 147 Urban/Suburban Arterial Segments 2U – 642 mi 3T – 160 mi 4U – 335 mi 4D – 269 mi 5T – 286 mi Urban/Suburban Arterial Intersections 3ST – 7,577 4ST – 2,535 3SG – 975 4SG – 2,762 Depending on approach and topic, a subset may be relevant as reference sites. High Rural Two- Lane: Intersections WA 2008– 2012 Segments MN 2003– 2009 Rural Multi-lane: Intersections OH 2009– 2011 Segments CA, TX, and OH 2009– 2011 Urban/ Suburban Arterials: Intersections OH 2009– 2011 Segments MN 2010– 2014 and OH 2007–2011 Could supplement with additional access - management- related variables. 2U = 2-lane undivided roadway segment, 3T = 2-lane roadway segment with TWLTL, 4U = 4-lane undivided roadway segment, 4D = 4-lane divided roadway segment, 5T = 4-lane roadway segment with TWLTL, 3ST = 3-legged minor stop-controlled intersection, 4ST = 4-legged minor stop-controlled intersection, 3SG = 3-legged signalized intersection, 4SG = 4-legged signalized intersection. Table 48. (Continued).

NCHRP Project 17-70 Intersection – Roundabouts only Area type Circulating lanes Number of legs Posted speed limit Inscribed circle diameter Major and minor road AADT 355 sites Depending on approach and topic, a subset may be relevant as reference sites. High CA, FL, KS, MI, MN, NC, NY, ON, PA, WA, WI (2000–2014 depending on site) Could supplement with additional access - management- related variables. Development of Crash Modification Factors (DCMF) B5 Pedestrian Countermeasure Crash Modification Factor Study Intersection – Signalized only Number of legs Number of lanes by type Presence of marked crosswalk One-way vs two-way operation Skewed vs non-skewed Type of left-turn phasing Major and minor AADT 143 sites 1,000+ sites used as reference sites in the study High Chicago, Toronto Data used in a before-after study of providing protected left - turn phasing Could supplement with additional access - management- related variables. Analysis of Right- In, Right-Out Commercial Driveway Safety, Operations, and Use of Channelization as Compliance Countermeasure. (Gorthy et al. 2017) Site – Right-in- right-out (RIRO) driveways Median presence Median type Presence of signage Driveway width AADT 9,000 driveways, 1,365 RIRO sites on 11 major business corridors Depending on approach and topic, a subset may be relevant as reference sites. Unknown SC 2011– 2014 Would require authors to share the data Safety Impacts of Directional Median Openings at Downstream U- Turn Locations. (Qi et al. 2016) Site – Downstream median opening for directional median openings Number of lanes Posted speed Peak-hour volumes Distance to U-turn opening 16 sites None in data Unknown Houston, TX 2007–2011 Would require authors to share the data (continued on next page)

Data Source Level of Data (corridor, site, intersection) Variables in Dataset Miles or Locations by Facility Type Availability of Reference Sites Data Quality Agencies That Provided Data Notes Safety Analysis of Driveway Characteristics Along Major Urban Arterial Corridors in South Carolina. (Stokes et al. 2016) Site - Driveways Driveway spacing Number of entry lanes Median type presence Driveway width Speed limit Driveway access control Presence of signalized intersections nearby Turning radius Angle Corner clearance Sight distance Driveway type Parking type AADT 9,000 driveways, 11 corridors Depending on approach and topic, a subset may be relevant as reference sites. Unknown SC 2010– 2012 Would require authors to share the data Safety Effects of Access Points near Signalized Intersections. (Jafari and Hummer 2013) Site - Driveways Lane configurations Corner clearance Speed limit Driveway angle Radius AADT Estimated driveway volumes 108 sites Depending on approach and topic, a subset may be relevant as reference sites. Unknown NC 2005– 2009 Would require authors to share the data Analysis of Safety Effects of Traffic, Geometric, and Access Parameters on Truck Arterial Corridors. (Sultana et al. 2014) Corridors AADT Truck AADT Median width Median opening density Corner clearance TWLTL length Left-turn bay length Driveway width Driveway throat width Driveway flare width Number of signals Driveway density by type Proportion of divided driveways Pavement quality Shoulder width 74 corridors, over 350 miles Unknown WI 2005–2009 Would require authors to share the data Table 48. (Continued).

Safety Evaluation of Turning Movement Restrictions at Stop-Controlled Intersections. FHWA DCMF Project. Intersections on urban 4- and 6- lane corridors RIRO vs full movement Number of lanes AADT Design speed Traffic control at downstream intersection Left-turn lane presence Lane width Shoulder width Distance from unsignalized intersection to downstream turn lane/location 333 unsignalized intersections 202 downstream intersections Depending on approach and topic, a subset may be relevant as reference sites. High CA Safety Effects of Corner Clearance at Signalized Intersections. FHWA DCMF Project. 4-legged signalized intersections with general corridor characteristics Corner clearance for mainline approaches Number of lanes Mainline lane width AADT Mainline posted speed Land use type Driveway density 275 signalized intersections Depending on approach and topic, a subset may be relevant as reference sites. High CA and Charlotte, NC Safety Evaluation of Restricted Crossing U -turn Intersection. FHWA DCMF Project. Intersections on suburban 4- or 6-lane roadways Angle Speed limits Number of through lanes Left-turn crossover lanes Right-turn lanes Presence of crosswalks Median widths AADT 11 signalized RCUTS and approximately 44 potential comparison sites Depending on approach and topic, a subset may be relevant as reference sites. High AL 2004– 2013 NC 2002– 2012 OH 2002– 2013 TX 2009– 2014 Safety Evaluation of Multiple Strategies at Signalized and Unsignalized Intersections. FHWA DCMF Project. Urban and rural signalized and unsignalized intersections on divided and undivided roads Area type Number of legs Number of lanes on major road AADT Signalized: 84 treatment sites 368 reference sites Unsignalized: 434 treatment sites and 568 reference sites Depending on approach and topic, a subset may be relevant as reference sites. High SC 2005– 2014 AADT data quality could be improved. Currently, only 1 year of AADT is available for all intersections. Safety Evaluation of Flashing Yellow Arrow at Signalized Intersections. FHWA DCMF Project. Intersection Major AADT Minor AADT Number of legs Left-turn phasing Through lanes on major road Median presence on major road Approaches with left-turn lanes 307 treated signals 438 untreated signals Depending on approach and topic, a subset may be relevant as reference sites. Unknown OK 2004– 2014 OR 2002– 2013 NV 2006– 2013 NC