National Academies Press: OpenBook

Evaluation of Safety Strategies at Signalized Intersections (2011)

Chapter: Chapter 5 - Safety Evaluation

« Previous: Chapter 4 - Prioritization of Strategies
Page 18
Suggested Citation:"Chapter 5 - Safety Evaluation." National Academies of Sciences, Engineering, and Medicine. 2011. Evaluation of Safety Strategies at Signalized Intersections. Washington, DC: The National Academies Press. doi: 10.17226/14573.
×
Page 18
Page 19
Suggested Citation:"Chapter 5 - Safety Evaluation." National Academies of Sciences, Engineering, and Medicine. 2011. Evaluation of Safety Strategies at Signalized Intersections. Washington, DC: The National Academies Press. doi: 10.17226/14573.
×
Page 19
Page 20
Suggested Citation:"Chapter 5 - Safety Evaluation." National Academies of Sciences, Engineering, and Medicine. 2011. Evaluation of Safety Strategies at Signalized Intersections. Washington, DC: The National Academies Press. doi: 10.17226/14573.
×
Page 20
Page 21
Suggested Citation:"Chapter 5 - Safety Evaluation." National Academies of Sciences, Engineering, and Medicine. 2011. Evaluation of Safety Strategies at Signalized Intersections. Washington, DC: The National Academies Press. doi: 10.17226/14573.
×
Page 21
Page 22
Suggested Citation:"Chapter 5 - Safety Evaluation." National Academies of Sciences, Engineering, and Medicine. 2011. Evaluation of Safety Strategies at Signalized Intersections. Washington, DC: The National Academies Press. doi: 10.17226/14573.
×
Page 22
Page 23
Suggested Citation:"Chapter 5 - Safety Evaluation." National Academies of Sciences, Engineering, and Medicine. 2011. Evaluation of Safety Strategies at Signalized Intersections. Washington, DC: The National Academies Press. doi: 10.17226/14573.
×
Page 23
Page 24
Suggested Citation:"Chapter 5 - Safety Evaluation." National Academies of Sciences, Engineering, and Medicine. 2011. Evaluation of Safety Strategies at Signalized Intersections. Washington, DC: The National Academies Press. doi: 10.17226/14573.
×
Page 24
Page 25
Suggested Citation:"Chapter 5 - Safety Evaluation." National Academies of Sciences, Engineering, and Medicine. 2011. Evaluation of Safety Strategies at Signalized Intersections. Washington, DC: The National Academies Press. doi: 10.17226/14573.
×
Page 25
Page 26
Suggested Citation:"Chapter 5 - Safety Evaluation." National Academies of Sciences, Engineering, and Medicine. 2011. Evaluation of Safety Strategies at Signalized Intersections. Washington, DC: The National Academies Press. doi: 10.17226/14573.
×
Page 26
Page 27
Suggested Citation:"Chapter 5 - Safety Evaluation." National Academies of Sciences, Engineering, and Medicine. 2011. Evaluation of Safety Strategies at Signalized Intersections. Washington, DC: The National Academies Press. doi: 10.17226/14573.
×
Page 27
Page 28
Suggested Citation:"Chapter 5 - Safety Evaluation." National Academies of Sciences, Engineering, and Medicine. 2011. Evaluation of Safety Strategies at Signalized Intersections. Washington, DC: The National Academies Press. doi: 10.17226/14573.
×
Page 28

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.

18 This Chapter provides a summary of the results of the eval- uation of the five treatments mentioned in Chapter 4. The first part of this Chapter gives an overview of the different evaluation methods that were used to develop the CMFs. Fol- lowing that is a summary of each evaluation that provides the description of the treatment, data used, methodology, and results. (Full details of each evaluation study are provided in a series of appendices which can be found online at http://apps. trb.org/cmsfeed/TRBNetProjectDisplay.asp?ProjectID=461.) Three evaluation methods were used in this study. The primary, and preferred one, is the EB before-after method, which is considered to be one of the best methods for con- ducting before-after studies in that it properly accounts for regression to the mean. The second method utilized is the comparison group before-after method. This method does not effectively address the bias due to regression to the mean, but is effective in accounting for other non-treatment effects such as those due to trends in crash reporting and changes in traffic volume. The third method is based on cross-sectional multiple regression models where the CMFs are derived based on the coefficients of variables in these models that pertain to the CMF. The cross-sectional regression models were used if the sample size for the EB evaluation was limited. A secondary objective of using cross-section models for some evaluations was to examine the comparability of before-after and cross- sectional studies, a subject of topical interest in CMF develop- ment, for which there is little research. An overview of these methods follows. More details can be found in a recent FHWA Guide entitled A Guide to Developing Quality Crash Modifica- tion Factors (Gross et al., 2010). Overview of Methods Before-After Analysis Using the Empirical Bayes Method The EB method properly accounts for regression to the mean bias in before-after studies. It also overcomes the difficulties of using crash rates in normalizing for volume differences between the before and after periods and properly accounts for differences in crash experience and reporting practice in amalgamating data and results from diverse jurisdictions. The EB method estimates the expected crashes that would have occurred in the after period (λ) and compares that with the number of reported crashes in the after period (π). The following steps are used to estimate λ: 1. Identify a reference group of untreated sites that is otherwise similar to the treatment group. 2. Use the reference group data to estimate safety performance functions (SPFs) (mathematical equations) that predict the number of crashes of different types as a function of traf- fic volumes and other site characteristics. Typically, SPFs are negative binomial regression models that are esti- mated using generalized linear modeling. 3. In estimating SPFs, calibrate annual SPF multipliers (time trend factors) to account for the temporal effects (e.g., variation in weather, demography, and crash reporting) on safety. 4. Use the SPFs, the annual SPF multipliers, and data on traffic volumes and site characteristics for each year in the before period for each treatment site to estimate the number of crashes that would be predicted in each year of the before period for each treatment site. 5. Use the predicted number of crashes in the before period (from the SPFs) and the observed crashes in the before period at each treatment site to estimate the EB-expected number of crashes in the before period in each site. The EB-expected crash frequency is then estimated to adjust for possible bias due to regression to the mean. 6. Estimate λ (expected crashes in the after period if the treatment had not been implemented) as the product of the EB-expected number of crashes in the before period and the sum of the annual SPF predictions for the after period divided by the sum of these predictions for the C H A P T E R 5 Safety Evaluation

19 before period (for each treatment site). The EB procedure also produces an estimate of the variance of λ. 7. The estimate of λ is then summed over all sites in a treat- ment group of interest and compared with the count of crashes during the after period in that group. The variance of λ is also summed over all sites in the strategy group. 8. These parameters (the summation of λ and its variance) are then used, along with the summation of crash counts after treatment, to estimate an effect of the treatment (CMF). The standard deviation of the CMF is also estimated, which makes it possible to determine if the CMF is statistically different from 1.0 for a specific level of significance. Before-After Analysis Using the Comparison Group Method This method does not account for regression to the mean but can be effective in accounting for other non-treatment effects such as those due to trends in crash reporting and changes in traffic volume. This method can make use of an untreated com- parison group of sites that are similar to the treatment sites used to estimate an SPF to account for changes in traffic volume and temporal trends in crash occurrence. Steps 1 through 3 that were discussed for the EB method could potentially be the same for the comparison group method as well. The departure from the EB method is that, instead of using steps 4 and 5 to estimate the expected number of crashes in the before period, the observed crashes in the before period is used for this esti- mate. This estimate could be biased if crashes are selected for treatment because of a randomly high observed crash count. Cross-Sectional Regression Models Cross-sectional studies derive CMFs by comparing the crash statistics from sites with and without the treatment. If it is possible to find sites that are similar to each other (apart from having or not having the treatment), then the CMF could be defined as the ratio of the average number of crashes in sites with the treatment to the average number of crashes in the sites without the treatment. In practice, it is very difficult to find sites that are similar to each other, and so regression models are typically used. The state of the art is to use negative binomial regression models where crash frequency is the dependent variable and the independent variables may include site characteristics including major and minor road AADT. The model coefficients are used to derive the CMFs. One problem with using cross-sectional models is that the differences in crashes between the sites with treatment and without treatment may be due to factors that were meas- ured and could not be included in the model, factors which could not be measured, or even factors that are unknown. Hence, at this time, the CMFs from cross-sectional models are not considered as reliable as CMFs derived from well-designed before-after studies, unless they can be corroborated with results from rigorous before-after studies. Further discussion of these methods is provided in the appendices that discuss the results of each evaluation in detail which can be found online at http://apps.trb.org/TRBNet ProjectDisplay.asp?ProjectID=461. Another resource is a re- cent publication from the Federal Highway Administration entitled A Guide to Developing Quality Crash Modification Factors (Gross, Persaud, and Lyon, 2010) that includes more information concerning different methods for developing CMFs. The rest of this chapter provides a summary of the re- sults obtained from each evaluation. Evaluation Summaries Installation of Dynamic Signal Warning Flashers Description of Treatment and Crash Types of Interest This analysis examines the safety impacts of installing dynamic signal warning flashers (DSWF) in advance of sig- nalized intersections. DSWF provides drivers with advance notice of the phase change. Specifically, the DSWF is linked to the signal, and flashers are actuated when the signal is about to change from green to yellow. The flashers are located in advance of the intersection and are actuated at a time when the driver would not be able to clear the intersection before the onset of the red phase. The basic objective was to estimate the change in crashes. Target crash types considered included the following: • All crash types (all severities); • Rear-end crashes (all severities); • Angle crashes (all severities); • Fatal and injury crashes (all crash types); and • Truck-related crashes (all severities). The change in crash frequency was analyzed by employing multiple methods using data gathered from three states. Appendix A provides the details associated with this evalua- tion along with example photographs. Data Used Departments of transportation in Nevada, Virginia, and North Carolina helped identify treatment sites (i.e., inter- sections where DSWF had been installed). They also provided geometric, traffic volume, and crash data. For Nevada, data from 1994 to 2008 were available, but only a subset of that data was used in order to avoid any major construction activity. For Virginia, data from 1998 to 2008 were available, and again

20 only a subset was used to avoid construction activity. For North Carolina, data from 1993 to 2009 were used. Methodology With respect to all the treatment sites in Nevada and most of the treatment sites in Virginia, it was discovered that the traffic signals and DSWF were installed at the same time. This observation had an important implication on the selection of an analytical method for this analysis. Since it would be diffi- cult to separate the effects of the signal installation from the effects of the DSWF installation, using before-after methods (e.g., comparison group method or EB method) for those sites would be problematic. In contrast, all of the treatment sites in North Carolina were already signalized when the DSWF were installed. Therefore, a before-after method could be employed with the North Carolina data without difficulty. Because of the issue with the timeframe for signal installations, a single method could not be employed for all three states. Instead, three methods were used: cross-sectional analysis, before-after with comparison group, and EB before-after. With respect to the Nevada data, a cross-sectional method was employed using two groups of sites: one group consist- ing of signalized intersections where DSWF were present and another group consisting of signalized intersections where DSWF were not present. In all, 261 site-years and 3,224 total crashes were included in this analysis. With respect to the Virginia data, two analytical methods were employed. A cross-sectional analysis was conducted with the Virginia data using two groups of sites: one group of sites consisting of signalized intersections where DSWF were present and another group consisting of signalized intersections where DSWF were not present. The Virginia cross-sectional analysis included 452 site-years and 1,201 total crashes. A before-after with comparison group method was also employed with the Virginia data with the goal of validating the results of the cross- sectional analysis. This analysis was possible because, for a sub- set of the treatment sites in Virginia, the DSWF installations occurred after the traffic signal installations. Another cross-sectional analysis was performed using a dataset which combined the Nevada and Virginia data. As with the individual state analyses, two groups were defined. One group consisted of sites in Nevada or Virginia where DSWF had been installed, and another group consisted of sites in Nevada or Virginia where DSWF had not been installed. With the North Carolina data, the DSWF were installed at intersections which were already signalized. Consequently, the problem of separating the effects of signal installation and DSWF installation was not present, and the state-of-the-art EB before-after method was used. The treatment group con- tained 14 sites, 1,000 total crashes in the before period, and 256 total crashes in the after period. The reference group consisted of 63 signalized intersections in North Carolina with 5,948 total crashes. Results The evaluation of DSWF utilized three analysis methods: cross-sectional, before-after with comparison group, and before-after with EB. The cross-sectional analyses for Nevada, Virginia, and the two states combined, show a consistent reduction in total crashes at intersections that had DSWF. The results from the before-after analyses validated these findings. The results also suggest that DSWF may help to reduce angle, injury, and heavy vehicle crashes, although the sample size was limited for many of the individual crash types. It was possible to investigate both angle and injury crashes using all three methods and the results consistently indicated a reduction in expected crashes with the presence of DSWF. The results were less consistent for rear-end crashes. The cross-sectional and comparison group analyses were similar, indicating a reduction in expected rear-end crashes with the presence of DSWF. However, the EB analysis indicated an increase in rear-end crashes. Note again that the cross-sectional and comparison group analyses were based on data from Nevada and Virginia, while the EB analysis was based on data from North Carolina. Multiple methods were used in this analysis of DSWF, resulting in multiple sets of CMFs. Of the various sets of CMFs produced in this analysis, the results of the combined cross-sectional analysis were ultimately deemed to be the most reliable. Table 5.1 presents the CMFs from the combined cross-sectional analysis, with the respective standard errors. It is important to note that the standard errors shown are ‘ideal’ standard errors, and the Highway Safety Manual recommends that these standard errors be increased by a factor of 2.0 for Total Crashes Rear-end Angle Injury & Fatal Heavy Vehicle CMF 0.814# 0.792# 0.745# 0.820# 0.956 Standard Error 0.062 0.079 0.086 0.083 0.177 Note: #Statistically significant at the 0.05 level (based on the ideal standard errors reported in this table) Table 5.1. Crash frequency CMFs (and standard errors) by crash type for installation of DSWF.

CMFs from cross-sectional regression models to account for the fact that results from cross-sectional models are not as reliable as those from well-designed before-after studies for estimating CMFs. The results seem to indicate that the dynamic signal warning flashers do provide a benefit with the largest percent reduction in angle crashes. The relatively large reduction in fatal and injury crashes is likely the greatest benefit of the dynamic sig- nal warning flashers in terms of overall safety. Future research could investigate the safety effects of the many variations of DSWF including roadside and overhead signs. Conversion of Signalized Intersections to Roundabouts Description of Treatment and Crash Types of Interest This analysis examined the safety impacts of converting signalized intersections to roundabouts. Roundabouts have the potential to reduce both the frequency and severity of crashes compared to a similar signalized intersection. The basic objective was to estimate the change in crashes. Target crash types considered included: • All crashes (all types and severities); • Property damage only crashes (all crash types); and • Fatal and injury crashes (all crash types). The change in total crash frequency was analyzed as well as the changes in different crash severities, recognizing that the treatment may have a different level of effect on the various severities. Appendix B provides the details associated with this evaluation. Data Used Geometric, traffic volume, and crash data for treatment sites were acquired from the States of Indiana (2003–2008); New York (3 years before and after treatment); Washington (2001–March 2009); Michigan (2000–2009); and North Car- olina (1999–2009) to facilitate the analysis. Data were also obtained from NCHRP Project 3-65 which was published as NCHRP Report 572: Roundabouts in the United States (Rodegerdts et al., 2007) where signalized intersections were replaced with roundabouts. NCHRP Project 3-65 provided data for 1 site in Florida, 3 sites in Colorado, 1 site in South Carolina, 2 sites in Maryland, and 1 site in Vermont for this analysis. A total of 28 sites were used in the evaluation (see Table 5.2). Data for reference sites (i.e., signalized intersections similar to those converted to roundabouts) were sought for use in developing the SPFs required for the EB methodology. Unfortunately, such data were difficult to obtain for all states in which treatment sites were identified. Reference sites were identified in Indiana, North Carolina, and New York. Crash, traffic volume, and geometric data were collected for the ref- erence group. The data from Indiana and North Carolina were used to directly calibrate SPFs for the two states. For all other locations, the SPFs previously used in NCHRP Project 3-65 were applied. In order to investigate the effect of approach speed on safety at the roundabouts, the research team attempted to obtain data from the different States regarding approach speed and/or speed limits. Data on approach speeds or speed limits were not available before the construction of the roundabouts. Speed limit and/or advisory speed data were obtained for the ‘after’ condition along the major road for each of the study sites. This was called “associated speed” and was based on the approach advisory speed when posted, and when it was not posted, based on the nearest upstream posted speed limit. Methodology The primary analysis methodology used was the EB before- after analysis as previously described. The evaluation analyzed the effects of the treatment on crash frequencies for different crash severities before and after the treatment. The EB analysis attempted to develop CMFs by severity (i.e., PDO vs. fatal/injury vs. total crashes). The reference sites from Indiana and North Carolina were used to develop SPFs for use in the EB before-after analysis. SPFs developed under a previous effort (NCHRP Project 3-65) were used for the other locations. In addition to treatment and reference sites, Indiana pro- vided data on additional intersections that were newly con- structed as roundabouts. It was not possible to include these sites in the before-after analysis because there was no before period. Instead, these data were used as part of a cross-sectional analysis employed to compare the safety performance of similar signalized intersections and roundabouts. The EB analysis was used to investigate the safety effects of converting signals to roundabouts, but the study was based Location Treatment Sites Colorado 3 1adirolF 3anaidnI Maryland 2 Michigan 2 New York 11 North Carolina 2 South Carolina 1 Vermont 1 Washington 2 82latoT Table 5.2. Number of sites for treatment group. 21

22 on a relatively small sample size. To further investigate the treatment, a cross-sectional study was employed, using neg- ative binomial regression models to analyze a larger sample of signalized intersections and roundabouts in Indiana and New York. The cross-sectional analysis was based on a total of 321 site-years, including 42 signalized intersections and 26 roundabouts. Several potential confounding factors were included in the cross-sectional analysis, including traffic volume, area type, number of approaches, and number of approach/roundabout lanes. Results The data collected and analyzed for this study show a general safety benefit for converting signalized intersections to round- abouts. The EB before-after analysis indicated a significant reduction in both total and injury crashes. A disaggregate analysis was also conducted to identify differential effects based on specific site characteristics (traffic volume, area type, num- ber of approaches, number of lanes, and associated speed). Regarding the effect on total crashes, the safety benefit of roundabouts appears to decrease as traffic volumes increase. The analysis also suggested that the safety benefit is larger for suburban than for urban conversions and for intersections with four approaches compared to those with three. There was no clear pattern regarding the effectiveness of the round- about with regard to ‘associated speed’ (as mentioned earlier, associated speed is the posted advisory speed or the nearest upstream posted speed limit on the major road during the ‘after’ period). Perhaps the most apparent and telling result of the disaggregate analysis is that the reduction in fatal and injury crashes is substantial and highly significant in all sce- narios. This is a result of the basic configuration of a round- about, where crossing-path and left-turn crashes are physically eliminated. While the study team employed the EB method to estimate the safety effects of converting signals to roundabouts, the study was based on a relatively small sample size. A cross-sectional analysis, employing negative binomial regression, was con- ducted to provide support for the EB analysis. Interaction terms were explored during the cross-sectional analysis to further investigate the relationship between traffic volume and the effect of installing roundabouts at signalized intersections. Interaction terms were significant in several of the cross- sectional models for total crashes, indicating differential effects for different volumes. The interaction term was insignificant in the injury-related models, confirming the sustained benefit across the range of traffic volumes. The results of the cross-sectional analysis are relatively con- sistent with, and corroborate, the results of the EB analysis. In particular, both the EB and cross-sectional analyses indicated that the effects of the treatment on total crashes may change as AADT changes. Specifically, with respect to total crashes, the safety benefit of roundabouts appears to decrease as traffic volumes increase. The two analysis methods also show a sub- stantial and sustained reduction in fatal and injury crashes for roundabouts across the range of traffic volumes. Based on the relative rigor of the EB method and the reason- ableness of the results, the recommended CMFs were taken from the EB analysis. Table 5.3 shows the CMFs and CMFunc- tions as applicable. For total crashes, the overall CMF was 0.792, but the CMF was found to increase (i.e., approach 1.0) with increasing AADT, and a CMFunction (0.00004*AADT + 0.303) was found to be appropriate. The CMFunction is applicable between a total intersection AADT of 5,300 and 43,000. Increasing the Change Interval Description of Treatment and Crash Types of Interest This analysis examined the safety impacts of modifying the change interval at signalized intersections. The change interval is the time allocated for the yellow and all red phases for a given approach. The basic objective was to estimate the change in crashes. Target crash types considered included: • All crashes (all types and severities); • Fatal and injury crashes (all crash types); • Angle crashes (all severities); and • Rear-end crashes (all severities). The change in total crash frequency was analyzed as well as the changes in different crash types and severities, recognizing that the treatment may have a different level of effect on the various types and severities. Appendix C provides the details associated with this evaluation. Condition Severity CMF / CMFunction All All 0.792 (0.050)# All 0.00004*AADT + 0.303 Injury and Fatal 0.342 (0.058)# 2-lane All 0.809 (0.061)# Injury and Fatal 0.288 (0.065)# 1-lane All 0.735 (0.086)# Injury and Fatal 0.451 (0.115)# Suburban All 0.576 (0.053)# Injury and Fatal 0.259 (0.066)# Urban All 1.150 (0.093) Injury and Fatal 0.445 (0.100)# 3 approaches All 1.066 (0.163) Injury and Fatal 0.370 (0.172)# 4 approaches All 0.759 (0.052)# Injury and Fatal 0.338 (0.061) # Note: #Statistically significant at the 0.05 level AADT is total intersection AADT *represents a product, i.e., 0.0004*AADT is the product of 0.0004 and AADT Table 5.3. Crash frequency CMFs (and standard deviations) by crash severity for converting signalized intersections to roundabouts.

Data Used Geometric, traffic volume, signal timing, and crash data for both treatment and reference sites were acquired from the States of California (1992–2002) and Maryland (1992–2002) to facilitate the analysis. Specifically, data were obtained in California from the cities of San Diego and San Francisco and in Maryland from the counties of Howard and Montgomery. The sites include data for two types of signalized intersections: (1) signalized intersections where the change interval was modified during the study period, and (2) signalized inter- sections where the change interval was not modified during the study period. If there were major changes to the geometry or operations during the study period, the sites were excluded. Methodology The primary analysis methodology used was the EB before- after analysis as previously described. The evaluation analyzed the effects of the treatment on crash frequencies for different crash types and severities before and after the treatment. Specifically, the EB analysis was employed to investigate five specific scenarios. Three scenarios were related to various combinations of increasing the yellow and all red time: • Increasing both the yellow and all red phases, • Increasing the all red phase only, and • Increasing the yellow phase only. Two other scenarios were investigated, comparing the total change interval to the ITE recommended practice (see Appendix C for a description of the ITE recommended prac- tice). In both cases, the before condition was represented by signalized intersections where the total change interval was less than the ITE recommended practice. The after period was represented by signalized intersections with the following characteristics: • Total change interval remains less than the ITE recom- mended practice and • Total change interval is greater than the ITE recommended practice. The analyses attempted to develop CMFs by severity (i.e., fatal/injury vs. total crashes) and by crash types (i.e., total, angle, and rear-end) for both States. The before-after analysis was based on a total of 31 treatment sites as noted in Table 5.4. Reference sites were identified in each jurisdiction to develop SPFs for use in the EB before-after analysis. In addition to treatment and reference sites, California and Maryland provided data on additional intersections that were signalized throughout the entire study period, but signal timing data were only available for a portion of the study period. These data were combined with the reference sites and data from the treatment sites in a cross-sectional analysis to investigate the individual yellow and all red phases with respect to the ITE recommended practice. The EB analysis was used to investigate the safety effects of modifications to the total change interval with respect to the ITE recommended practice. Due to a relatively small sample size, it was not possible to investigate the individual yellow and all red phases with respect to the ITE recommended practice, using the EB method. Instead, a cross-sectional study was employed, using negative binomial regression models to ana- lyze a larger sample of signalized intersections with various combinations of yellow and all red phases. The cross-sectional analysis was based on a total of 916 site-years where the specific yellow and all red time were known for each year. Results In discussing the results, it should be noted that the mod- ifications to the yellow and all red time were not equivalent for all sites. This applies to both the existing conditions and the increase in the yellow and/or all red intervals. For example, several of the intersections did not include an all red phase in the before condition. For sites where both the yellow and all red time were increased, the average increases in the yellow and all red times were 0.8 seconds (minimum of 0.5 seconds and maximum of 1.6 seconds) and 1.2 seconds (minimum of 1.0 second and maximum of 2.0 seconds), respectively. For sites where only the yellow interval was increased, the increase in yellow time was 1.0 second in all the sites. For sites where only the all red interval was increased, the average increase in the all red time was 1.1 seconds (minimum of 1.0 second and maximum of 2.0 seconds). For sites where the total change interval was increased, but still less than the ITE recommended practice, the average increase was 0.9 seconds (minimum was 0 seconds and the maximum was 1.5 seconds). For sites where the total change interval was increased and exceeded the ITE recommended practice, the average increase was 1.6 seconds (minimum was 1.0 second and maximum was 3.0 seconds). Based on the rigor of the EB method, and the generally insignificant results of the cross-sectional analysis, the rec- ommended CMFs were taken from the EB analysis. Location Treatment Sites Reference Sites Howard County, MD 2 29 Montgomery County, MD 6 38 San Diego, CA 16 36 San Francisco, CA 7 32 53113latoT Table 5.4. Number of sites for treatment and reference groups. 23

Treatment Crash Type Severity CMF (S.E. of CMF) Average Increase in Total Change Interval (min, max) Number of All Red Intervals = 0 Before Treatment Increase Change Interval (< ITE) (12 sites) All All 0.728 (0.077)# 0.9 (0, 1.5) 11 All Injury & Fatal 0.662 (0.099)# Rear-end All 0.848 (0.142) Angle All 0.840 (0.195) Increase Change Interval (> ITE) (15 sites) All All 0.922 (0.089) 1.6 (1.0, 3.0) 10 All Injury & Fatal 0.937 (0.114) Rear-end All 0.643 (0.130)# Angle All 1.068 (0.156) Note: #Statistically significant at the 0.05 level Table 5.6. Crash frequency CMFs (and standard errors) by crash type for increasing the change interval. 24 The EB before-after analyses indicated a significant reduction in total, injury, and rear-end crashes under various scenarios. Specifically, the EB analysis indicated a statistically significant reduction (at the 0.05 level) in total crashes as a result of (1) increasing the all red phase only, and (2) increasing the total change interval to be less than the ITE recommended practice. Injury crashes were significantly reduced as a result of increasing the total change interval to be less than the ITE recommended practice. Rear-end crashes were significantly reduced as a result of increasing the total change interval to be greater than the ITE recommended practice. The change in angle crashes was statistically insignificant under all scenarios investigated. Table 5.5 shows the CMFs and standard errors for total, injury, rear-end, and angle crashes as they relate to increasing the yellow and/or all red intervals. Table 5.6 shows similar results for increasing the total change interval. Each table also indicates the average increase in the respective interval, the applicable range of values, and the number of sites without an all red phase in the before period. It is important to note that the number of sites in this evaluation was limited, and hence the results should be treated with due caution. Change Left-Turn Phasing (From Permissive to Protected-Permissive) Description of Treatment and Crash Types of Interest The objective was to estimate the general safety effects of changing from permissive to protected-permissive phasing at signalized intersection approaches. Additionally, a particular goal was to investigate the effects on non-left-turn related crash types and look at the effects of traffic volume, left-turn volume, and number of opposing lanes on the estimated change in crashes. Treatment (Number of sites) Crash Type Severity CMF (S.E. of CMF) Average Increase in Yellow Interval (min, max) Average Increase in All Red Interval (min, max) Number of All Red Intervals = 0 Before Treatment Increase Yellow and All Red (11 sites) All All 0.991 (0.146) 0.8 (0.5, 1.6) 1.2 (1.0, 2.0) 11 All Injury & Fatal 1.020 (0.156) Rear-end All 1.117 (0.288) Angle All 0.961 (0.217) Increase Yellow Only (5 sites) All All 1.141 (0.177) 1.0 (1.0, 1.0) -- 1 All Injury & Fatal 1.073 (0.216) Rear-end All 0.934 (0.237) Angle All 1.076 (0.297) Increase All Red Only (14 sites) All All 0.798 (0.074)# -- 1.1 (1.0, 2.0) 10 All Injury & Fatal 0.863 (0.114) Rear-end All 0.804 (0.135) Angle All 0.966 (0.164) Note: #Statistically significant at the 0.05 level Table 5.5. Crash frequency CMFs (and standard errors) by crash type for increasing the yellow and/or all red interval.

The site types of interest were signalized intersections with left-turn lanes in either urban or rural environments, which have been converted to protected-permissive for at least part of the daily operation. The following crash types were of interest: • Total crashes; • Injury crashes; • Left-turn crashes; • Left-turn opposing through crashes (crashes involving a left-turn vehicle and a through vehicle from the oppos- ing approach); and • Rear-end crashes. Appendix D provides the details of this evaluation. Data were acquired from the City of Toronto, Canada, and urban areas in North Carolina, for both treated and untreated signalized intersections. Data from Toronto The City maintains a database of signalized intersections including many variables related to geometry (e.g., number of lanes by type by approach), traffic volumes, and crash data. Volume and crash data from 1999 to 2007 were collected. This database was augmented by querying the crash data for specific crash types and adding left-turn AADTs. A separate database of intersection approaches was also created as it was desired to evaluate left-turn protection improvements at both the intersection-level and approach-level. Intersections at which only one approach had an improvement in left-turn protection were used for the approach-level analysis. Treated sites were identified in a two-step process. First, an electronic file of work orders for signalized intersections was scanned to identify sites where a change in left-turn phasing was made. Using this list, a subsequent search of hard copy signal timing reports for these sites identified those where the left-turn phasing on at least one approach was changed to either protected-permissive or fully protected at any time of day. The group of 59 intersection level and 46 approach level treatment sites represented a range of before and after condi- tions with regard to left-turn phasing options. Hence, sites were categorized based on the predominant phasing system. A reference group of untreated signalized intersections was identified to match the treatment sites based on site characteristics, including number of approaches, presence of left-turn lanes, and traffic volumes. Data from North Carolina In North Carolina, data were available for 19 four-leg inter- sections that experienced a change in left-turn phasing on at least one leg of the intersection. All these 19 sites were in urban areas. The change in phasing was one of the following three categories: • From Permissive to Protected-Permissive (12 intersections); • From Permissive or Protected-Permissive to Protected (5 intersections); and • From Protected to Permissive or Protected-Permissive on at least 2 legs (2 intersections). Since the number of intersections in the last two categories is very limited, results are provided here only for the first category of sites, i.e., for intersections where the phasing was changed from permissive to protected-permissive phasing in at least one leg of the intersection. All the treatment locations had a left-turn lane on the major legs of the intersection. Unlike Toronto, crash data by approach were not available in North Carolina without a manual review of crash reports. So, in North Carolina the analysis was focused at the inter- section level. Methodology The methodology applied was the empirical Bayes (EB) before-after study, which was described at the beginning of this chapter. Further details about the methodology are pro- vided in Appendix D. A number of SPFs were calibrated as follows: • SPFs were calibrated separately for Total, Injury, Left-turn, Left-turn-opposing through, and Rear-end crashes. • SPFs at the intersection-level and approach-level were sep- arately developed for Toronto. For North Carolina, SPFs were estimated at the intersection-level. • For the City of Toronto, separate models were also developed for intersections without and with one-way roads. Results The results are shown in Tables 5.7 and 5.8. Approach level results are based on data from Toronto. Intersection level results are based on data from both Toronto and North Crash Type CMF (s.e.) All 1.077 (0.037) # Injury and Fatal 1.150 (0.056) # LTOPP 0.776 (0.098) # Rear end 1.103 (0.118) Note: #Statistically significant at the 0.05 level Table 5.7. Approach level results (Toronto). 25

26 Carolina (all intersections were four-leg). Intersection level results are provided for two categories of intersections: inter- sections where only 1 approach was treated and intersections where more than 1 approach was treated. Among the 21 inter- sections where more than 1 approach was treated, 17 of them had 2 approaches treated, 2 of them had 3 approaches treated, and 2 of them had 4 approaches treated. At both intersection and approach levels, the results indicate substantial benefits for the target crash type, left-turn oppos- ing involving a left-turn vehicle and a through vehicle from the opposing approach (LTOPP). As expected, the benefit at the intersection level is greater at intersections where more than one approach is treated. One of the fundamental questions the study was expected to answer was the extent to which the decrease in target crashes may be offset by a compensating increase in a non-target crash type such as rear-end. At both the intersection and approach levels, there were small percentage increases in rear-end crashes. The actual (rather than percentage) increase in rear-end crashes was of the order of 60–75% of the decrease in left-turn oppos- ing crashes. Disaggregation of the effects by AADT, either total entering or left turn, did not reveal any trend. This may be because the intersections did not have a wide enough dis- tribution of these variables. In summary, it may be concluded that in estimating the net safety benefit of left-turn protection, consideration must be given to the increase in non-target crashes as well as the decrease in target crashes. It is recommended that the intersection level results for Toronto and North Carolina be used to refine the current CMF for changing from permissive to protected- permissive in the Highway Safety Manual. Further research could investigate the specific safety effects of changing left- turn phasing during particular times of day (e.g., peak versus off-peak) and days of the week (e.g., weekday versus weekend). Another area of research is to investigate the effect of combined left-turn treatments: adding a left-turn lane and changing the left-turn phase at the same time. There were a few sites in North Carolina where such combined treatments were imple- mented, but they were not sufficient to conduct an evaluation. Installation of Flashing Yellow Arrow for Permissive Left Turns Description of Treatment and Crash Types of Interest The objective was to evaluate the safety impacts due to the installation of flashing yellow arrow (FYA) for permissive left-turn movements. The intent of the flashing yellow arrow is to avoid the confusion for drivers turning left on a permis- sive circular green signal indication who may assume that the left turn has the right of way over opposing traffic, especially under some geometric conditions. The following primary tar- get crash types were considered: • Total intersection crashes; • Total left-turn crashes; and • Total left-turn crashes from the FYA treated approach (this crash type was examined in Washington and Oregon, but not in North Carolina). Appendix E provides the details of this evaluation. Data Used The data included 5 locations in Kennewick, Washington, 34 locations from cities in Oregon, and 16 locations from urban areas in North Carolina. In Kennewick, FYA was introduced in these five locations between 2004 and 2006. Four of these locations had protected-permissive phasing before FYA was introduced and one location had permissive phasing before FYA was introduced. The City of Kennewick provided many variables related to geometry (e.g., number of lanes by type and approach), traffic volumes in the form of major and minor road AADTs, peak hour left-turn movements, and crash data. In Oregon, the city of Beaverton provide data for 15 sites, the city of Gresham provided data for 6 sites, the city of Oregon City provided data for 3 sites, and the city of Portland provided data for 10 sites with FYA. Twenty-four of these locations had protected phasing before the FYA was introduced, 3 of them had permissive phasing, 3 of them had protected-permissive phasing, and 4 had prohibited left turns. The four cities in Oregon were able to provide many variables related to geometry (e.g., number of lanes by type by approach) and crash data. Left-turn volumes were not available for the Oregon locations. In North Carolina, in all 16 intersections that were evalu- ated, flashing yellow arrow (FYA) was introduced in two Crash Type Grouping No. Sites CMF (s.e.) All All sites 71 1.033 (0.023) 1 treated approach 50 1.085 (0.028) # >1 treated approach 21 0.945 (0.040) Injury and Fatal All sites 71 0.958 (0.037) 1 treated approach 50 1.005 (0.045) >1 treated approach 21 0.878 (0.062) # LTOPP All sites 71 0.858 (0.056) # 1 treated approach 50 0.919 (0.069) >1 treated approach 21 0.762 (0.088) # Rear end All sites 71 1.063 (0.038) 1 treated approach 50 1.091 (0.046) # >1 treated approach 21 1.021 (0.062) Note: #Statistically significant at the 0.05 level Table 5.8. Intersection level results (Toronto and North Carolina combined).

out of the four legs. The changes were divided into the fol- lowing three categories: • Change from protected phasing to FYA protected-permissive in 2 legs of the intersection (5 intersections); • Change from doghouse (conventional protected-permissive) to FYA protected-permissive in 1 leg and from permissive to FYA protected-permissive in another leg (5 intersections); and • Change from doghouse (conventional protected-permissive) to FYA protected-permissive in 2 legs of the intersection (6 intersections). In North Carolina, turning volumes were not available at the treatment or reference sites. However, data on major and minor road AADT were available for the treatment and reference sites. Methodology For the cities in Washington and Oregon, data on reference sites was limited in most of the jurisdictions, and hence the EB methodology could not be applied with the required rigor. The cities did indicate that the sites were not selected based on crash history, but some evidence of an absence of regression- to-the-mean was still desired. The investigation of potential regression-to-the-mean involved aggregating the crash data over all treatment sites and plotting the totals for each year before treatment (e.g., 1 year before treatment, 2 years before treatment, 3 years before treatment, etc.). This test was conducted for each city separately, and for each, it was con- cluded that there was no evidence for regression-to-the-mean notwithstanding the natural randomness of crash counts. The methodology applied combined some aspects of the EB and Comparison Group approaches. Adjustments for changes in AADT were done by using an SPF calibrated for Kennewich, WA, which had sufficient sites for this purpose, and then dividing the SPF estimate using the after period AADT by the SPF estimate using the before period AADT. The adjustment for time trends was determined using a group of comparison sites by dividing the sum of SPF predictions per year for the after periods by the sum of SPF predictions per year in the before period for the comparison group. In North Carolina, the state of the art EB method could be applied. Safety performance functions were estimated using a reference group of 49 intersections in North Carolina. Further detail about the methodology is provided in Appendix E. Results Crash Modification Factors are provided in Table 5.9 for total intersection crashes and total intersection left-turn crashes (the common crash type investigated in the 3 states). Results are provided for three categories of changes depending on the left-turn phasing of the converted legs before FYA was introduced: • Intersections where the converted legs had either permis- sive or protected-permissive phasing in the before period, and at least one of the legs had permissive phasing. This group includes 9 four-leg intersections (total of 36 legs). A total of 20 legs were treated with FYA: 15 of the treated legs had permissive phasing in the before period while 5 of the treated legs had protective-permissive phasing in the before period. • Intersections where the converted legs only had protected- permissive phasing in the before period. This group included 1 3-leg and 12 4-leg intersections (total of 51 legs). A total of 27 legs were treated with FYA; all of them had protected- permissive phasing in the before period. Left-Turn Phasing Before (sites) (legs treated) Crash Type CMF (S.E.) Permissive or combination of permissive and protected-permissive (at least 1 converted leg was permissive in the before period) (9 sites) (20 legs treated) Total Intersection Crashes 0.753 (0.094) # Total Intersection Left-Turn Crashes 0.635 (0.126) # Protected-Permissive (all converted legs had protected-permissive in the before period) (13 sites) (27 legs treated) Total Intersection Crashes 0.922 (0.104) Total Intersection Left-Turn Crashes 0.806 (0.146) Protected (all converted legs had protected in the before period) (29 sites) (56 legs treated) Total Intersection Crashes 1.338 (0.097) # Total Intersection Left-Turn Crashes 2.242 (0.276) # Note: #Statistically significant at the 0.05 level Table 5.9. CMFs and standard errors for flashing yellow arrow installation. 27

• Intersections where the converted legs only had protected only phasing in the before period. This group included 5 3-leg intersections and 24 4-leg intersections (total of 111 legs). A total of 56 legs were treated with FYA; all of them had protected only phasing in the before period. Intersections in the first group experienced reductions in total intersection crashes and total intersection left-turn crashes that were statistically significant at the 0.05 level. Intersections in the second group experienced a smaller reduction that was not statistically significant at the 0.05 level. As expected, on the basis of individual results and those in Noyce et al. (2007), intersections in the third group (with protected only phasing in the before period) experienced significant increases in total and left-turn crashes. As Noyce et al. commented, the change in signal phasing may have had a more significant impact on safety than the change to FYA permissive indication. Collec- tively, these results indicate that the largest benefit can be found at sites where at least one of the converted legs had per- missive only operation before the FYA was implemented with protected-permissive operation. It is important to note that the number of sites in the first two groups was limited, and hence the individual results should be treated with due caution. Most of the sites had 2 legs that were converted (except for the few 3-leg intersections in the sample). So, it was not possi- ble to specifically investigate the relationship between the num- ber of legs that are treated and the associated safety benefits for left-turn crashes. This could be an area for future research. Another area of future research is an investigation into the effect of left-turn volume and opposing through volume. 28

Next: Chapter 6 - Compilation of CMFs »
Evaluation of Safety Strategies at Signalized Intersections Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB’s National Cooperative Highway Research Program (NCHRP) Report 705: Evaluation of Safety Strategies at Signalized Intersections explores crash modification factors (CMFs) for safety strategies at signalized intersections. CMFs are a tool for quickly estimating the impact of safety improvements.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!