National Academies Press: OpenBook

Safety Evaluation of Permanent Raised Pavement Markers (2004)

Chapter: Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices

« Previous: Chapter 1 - Introduction
Page 4
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 4
Page 5
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 5
Page 6
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 6
Page 7
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 7
Page 8
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 8
Page 9
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 9
Page 10
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 10
Page 11
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 11
Page 12
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 12
Page 13
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 13
Page 14
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 14
Page 15
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 15
Page 16
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 16
Page 17
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 17
Page 18
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 18
Page 19
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 19
Page 20
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 20
Page 21
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 21
Page 22
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 22
Page 23
Suggested Citation:"Chapter 2 - Review of PRPM-Related Literature and Jurisdictional Practices." National Academies of Sciences, Engineering, and Medicine. 2004. Safety Evaluation of Permanent Raised Pavement Markers. Washington, DC: The National Academies Press. doi: 10.17226/13724.
×
Page 23

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.

4CHAPTER 2 REVIEW OF PRPM-RELATED LITERATURE AND JURISDICTIONAL PRACTICES This chapter summarizes (1) the findings of a literature review of studies related to PRPMs and (2) the state of the practice according to state departments of transportation (DOTs) that responded to the iTRANS survey. Three sections make up this chapter. First, an overview of current PRPM guidelines and practices is provided. This overview is followed by a section that critically reviews the knowledge about the safety effect of PRPMs. The section also assesses method- ological problems and issues arising from past research. This assessment sets the stage for the design of the current evalua- tion study. The third section reviews the human factors issues studied in past research efforts that may become relevant to the interpretation of the study results and formulation of guidelines for PRPM implementation. 2.1 OVERVIEW OF PRPM CURRENT PRACTICES PRPMs were developed to provide delineation over a wider range of environmental conditions than could be achieved with standard pavement marking materials. Retroreflective PRPMs provide a clear, definitive outline of pavement markings even under adverse visibility conditions such as rain, fog, and dark- ness. Standard paint lines are ineffective during rainy condi- tions because rain often accumulates on the painted markings, thus reducing the retroreflectivity of the paint, whereas the raised pavement markers stand above the pooled water (2). Nonsnowplowable PRPMs are also effective in providing a “wake up call” for the driver who wanders out of the travel lane. The wake up call is created by the vehicle vibration and audible tone when crossing over the PRPMs. Currently, there are many types and models of pavement markers on the mar- ket, including both retroreflective and nonretroreflective types. PRPMs are developed with a variety of configurations and characteristics. Some markers are wedge shaped, and some are round or oval. Markers are available with and without replace- able retroreflective inserts, and the variations are numerous. Commercially available markers vary in all aspects, such as size, shape, composition, and capabilities. Snowplowable markers have also been developed to reduce the vulnerabil- ity of the marker to snowplow activity. In general, there are two main types of PRPMs: retroreflective and nonretro- reflective. In addition, there are two subcategories of retro- reflective PRPMs: conventional raised nonsnowplowable pavement markers and snowplowable pavement markers. Snowplowable markers can be either raised or recessed. 2.1.1 Nonretroreflective PRPMs Nonretroreflective raised pavement markers, such as con- vex buttons, are made of plastic, ceramic, or aluminum. Non- retroreflective raised pavement markers are often used, in conjunction with retroreflective PRPMs, as an alternative to painted markings. The “Manual on Uniform Traffic Control Devices” (MUTCD) (3) provides guidelines on the spacing requirements of nonretroreflective PRPMs when used as a replacement to painted markings. The “Roadway Delineation Practices Handbook” (2) indicates that the nonretroreflective PRPMs will provide daytime visibility and the retroreflective PRPMs the nighttime visibility. 2.1.2 Retroreflective PRPMs There are two types of retroreflectors: prismatic (also called cube-corner) and spherical lens (4). The marker that is most commonly installed today is the wedge-shaped, cube-corner retroreflector. Prismatic retroreflectors are manufactured with different face designs, and spherical lens retroreflectors are manufactured with different glass bead types. Changing either the face design or the glass bead type in retroreflectors can be expected to give different results for daytime and night- time delineation. References such as ASTM F923-00 (4), the “Roadway Delineation Practices Handbook” (2), and the “Guidelines for the Use of Raised Pavement Markers” (5) provide guidance in the selection of appropriate markers. The American Society for Testing and Materials (ASTM) has produced the following standard specifications: • ASTM D 4383, “Standard Specification for Plowable, Raised Retroreflective Pavement Markers” (6), and • ASTM D 4280, “Standard Specification for Extended Life Type, Nonplowable, Prismatic, Raised Retroreflec- tive Pavement Markers” (7).

These specifications provide the performance require- ments for pavement markers in terms of the coefficient of luminous intensity before abrasion, abrasion resistance, color, lens impact strength, adhesive bond strength, compressive strength, and ramp hardness of holders. These specifications also provide guidelines on how to test the performance of pavement markers. Many states have developed special provisions and standard specifications for pavement markers that are generally based on the above ASTM specifications. While the ASTM specifi- cations provide minimum and maximum values, state guide- lines tend to identify more precise values that fall within the ranges of acceptable values specified in the ASTM guidelines. For example, ASTM D 4383 specifies that the installed height of the casting shall not exceed 0.43 in. (10.9 mm) and shall not be less than 0.06 in. (1.5 mm) above the road sur- face, while the guidelines for the state of Illinois specify that the height should be 0.3 in. (7.6 mm) and the guidelines for the state of Maryland specify a maximum height of 0.25 in. (6.4 mm). 2.1.2.1 Snowplowable Pavement Markers In the United States, there are two types of snowplowable pavement markers: raised and recessed. According to the iTRANS state surveys and literature reviews, recessed mark- ers are not as popular as raised markers are. Some of the states that have installed or are currently installing recessed mark- ers are Kansas, Maine, Maryland, Oregon, Virginia, West Vir- ginia, and Pennsylvania. These states, with the exception of Oregon, also install raised snowplowable pavement markers. Illinois, Indiana, Massachusetts, Michigan, New York, Ohio, and Wisconsin almost exclusively use raised snowplowable markers. Hofmann and Dunning (8) found that, although recessed snowplowable markers last on average 12 months longer than raised snowplowable pavement markers, they do not perform as well as raised markers. This finding confirms Endres’s (9) conclusion that raised pavement markers out-perform recessed markers under dry and wet weather conditions. A variety of problems are associated with recessed markers because the collection of debris, rain, and snow in the recessed slots obscure the reflective surface of the markers. Pigman and Agent (10) evaluated the performance of recessed snowplow- able markers by observing the marker’s visibility during snow and ice conditions. It was found that following snow- plow operations, the groove retained snow and ice. How- ever, because of the passing traffic, the snow and ice melted and the water was swept away in a short period of time. Pigman and Agent observed that vehicle tires cleansed the top third of the marker, but the bottom portion remained obscured. It was concluded that although nighttime visibility was reduced, the recessed markers remained visible. 5 Some states have evaluated the performance of recessed markers. The state of Maine ceased the installation of recessed markers because when the recessed grooves become filled with snow and ice, the markers are ineffective. Investigations by the Pennsylvania DOT (PennDOT) found that recessed markers on downgrades are not as visible as recessed mark- ers on inclines if water accumulates in the recessed slots. As a result, PennDOT has decided to stop the installation of recessed markers on its roadways. 2.1.2.2 Nonsnowplowable Pavement Markers Raised nonsnowplowable markers are used extensively in states where snowfall is not a concern, such as Texas and Cal- ifornia. Other states—such as Illinois, New Jersey, Oregon, Michigan, Maryland, and Massachusetts—use only snow- plowable pavement markers. 2.1.3 Implementation Criteria and Maintenance Procedures States extensively use the MUTCD (3) and FHWA’s “Roadway Delineation Practices Handbook” (2) as guides for the implementation of PRPMs. The MUTCD mainly provides guidelines on the desired spacing of PRPMs, while the “Road- way Delineation Practices Handbook” also provides general guidelines on PRPM colors, materials, installation, and main- tenance procedures. The “Roadway Delineation Practices Handbook” provides guidelines on the desired layout of PRPMs for various roadway infrastructure elements (e.g., curves, intersections, and tangents ramps) on different road- way types (e.g., two-lane roadways, four-lane undivided road- ways, and four-lane divided roadways). According to these guidelines, the spacing between consecutive PRPMs on tan- gents should be 80 ft (24 m). For horizontal curves between 3 and 15 degrees, a spacing of 40 ft (12 m) is recommended. For curves greater than 15 degrees, the recommended spac- ing is 20 ft (6 m). It is not recommended that centerline and edgeline PRPMs be used together because this may create confusion on some sharp curves. Most states, in accordance with these guidelines, install one two-way yellow marker on the centerline of two-lane roadways only. In some states, such as Illinois and Pennsylvania, a group of two markers can be used on the centerline of high-volume, high-speed, two- lane roads. On divided multilane facilities, the most common practice in the iTRANS states surveyed during this current research study is to install one-way white PRPMs on the lane lines only. An exception is New Jersey, where PRPMs are also installed on the left edgelines of multilane facilities. States have developed PRPM installation criteria. In the states of Ohio, Texas, and California, PRPMs are installed non- selectively on all state-maintained highways. Other states— such as Maryland, Massachusetts, Wisconsin, Pennsylvania, Illinois, Indiana and Kansas—have a combination of selective

and nonselective implementation practices. PRPMs are imple- mented nonselectively on certain roadway types, such as free- ways, and selectively on other roadway types on the basis of one or more of the following parameters: • Roadway type, • Traffic volume, • Illumination, • Safety record, • Speed limits, and • Horizontal curves. For example, Maryland implements PRPMs nonselec- tively on all Interstate highways and other freeways. Mary- land, Massachusetts, and Wisconsin use the speed limit of a roadway as a primary criterion for deciding where to imple- ment PRPMs. In Maryland, PRPMs are implemented on all two-lane roadways that have a speed limit exceeding 45 mph (72 km/h), irrespective of the traffic volume. In Massachu- setts, PRPMs are installed on undivided roadways that have a speed limit of 50 mph (80 km/h) or greater. In Wisconsin, PRPMs are installed on all roadways that have a speed limit of more than 65 mph (100 km/h), which includes all multi- lane freeway facilities. Missouri, Pennsylvania, and Massachusetts implement PRPMs on all freeways. Michigan’s PRPM guidelines rec- ommend implementation on all freeways that lack roadway illumination. The criteria for implementing PRPMs in Illinois, Indiana, and Kansas relate to traffic volume thresholds for different roadway types. PRPMs are only installed on roadways where the average daily traffic (ADT) volumes exceed these thresh- olds. Table 2-1 provides a summary of the traffic volume thresholds for different roadway types. The majority of surveyed states implement PRPMs at locations with actual or potentially poor safety records. In Maryland, PRPMs are implemented where the crash rate for “correctable” guidance-related crashes is significantly higher than the statewide average on similar road types. In Indiana, site selection for the implementation of PRPMs is based pri- marily on the need for additional alignment delineation in areas of frequent inclement weather (e.g., fog, smoke, and 6 rain); low roadway illumination; and evidence of vehicles leaving the roadway, such as excessive wear of pavement markings or excessive skid marks. In Michigan, PRPMs are installed only on nonfreeways where there is a concentration of crashes and only after other countermeasures such as sign- ing, pavement markings, and roadside delineation (e.g., chevrons and post-mounted delineators) have been unsuc- cessful in improving the safety of the locations. Illinois and Maryland install PRPMs at horizontal curves where it is necessary for motorists to decrease their travel speed by more than 10 mph (16 km/h) in order to traverse the curve safely. Some states implement PRPMs at other cross-section ele- ments. For example, Illinois installs PRPMs at lane reduc- tion transitions; freeway gores; rural left-turn lanes; and two-way, left-turn lanes. Maryland has detailed standard design drawings for PRPM installations at one-lane bridges; intersection approaches; two-way, left-turn lanes; left-turn lanes; acceleration lanes; deceleration lanes; and lane transitions. One of the primary maintenance problems with retro- reflective PRPMs is maintaining the reflectivity level. The reflectivity retention of retroreflective PRPMs tends to depend mostly on cumulative vehicular exposure since the time of installation (11). A study by Ullman (12) evaluated several models of corner-cube reflectors for factors such as volume of vehicle exposure, degradation in reflectivity, damage, and missing percentages. The “Roadway Delineation Practices Handbook” (2) states that it is difficult to precisely predict the service life of retroreflective PRPMs. In response to the iTRANS state practices survey, some states provided information on their PRPM maintenance practices. Pennsylvania and Ohio replace PRPM lenses on a fixed 2-year and 3-year cycle, respectively. In some states, the replacement cycle depends on the roadway type and traf- fic volume. Table 2-2 shows the PRPM replacement cycle for Indiana. Texas provides guidelines for when to schedule the maintenance of PRPMs based on the results of a nighttime test inspection (Table 2-3). The replacement cycle of PRPMs in Texas, based on ADT volumes, is summarized in Table 2-4. Colorado and Iowa removed all existing PRPMs and interrupted any future installations because of the high maintenance costs. State Guidelines for rural two- lane roadways Guidelines for multilane roadways Illinois ADT > 2,500 veh/day ADT > 10,000 veh/day Indiana ADT > 2,500 veh/day ADT > 6,000 veh/day Kansas ADT > 3000 veh/day and TADT > 450 veh/day ADT = Average daily traffic (both directions). TADT = Truck average daily traffic. TABLE 2-1 PRPM guidelines based on traffic volume for different roadway types (source: iTRANS state practices survey)

2.2 REVIEW AND ASSESSMENT OF KNOWLEDGE ABOUT THE SAFETY EFFECT OF PRPMS This section critically reviews literature before summariz- ing any methodological problems arising from past research. The literature review focuses on the relatively few studies that have been conducted from about 1980 to date, since older studies are less likely to be relevant in terms of findings and methodologies employed. 2.2.1 Review of Literature Seven evaluation studies were reviewed and are summa- rized herein. The first study to be reviewed was undertaken by Wright et al. (13). This study evaluated the safety effects of reflec- tive raised pavement markers in Georgia. From 1976 to 1978, the Georgia DOT installed reflectorized pavement markers (both raised and recessed markers) on the centerlines of 662 horizontal curves, all of which were in excess of 6 degrees of curvature. At some locations, warning signs, chevron mark- ers, or other delineation devices that were intended to pro- vide guidance to drivers were also installed. These additional devices may have affected the analysis’s results. For each curve studied, the location, length, degree of curvature, year of installation, ADT by year, and annual crash frequency by type (single-vehicle or other) and time of day (day or night; daytime: 6:00 a.m. to 5:59 p.m.) were collected. Locations were monitored 200 ft (61 m) in both directions beyond the curve in the belief that curve-related, single-vehicle crashes often take place beyond the end points of curves. 7 The study examined the change in nighttime crashes (from 6:00 p.m. to 5:59 a.m.) and used daytime crashes at the same sites as a control group. In the crash data from 1975 to 1980, there were 223 before-installation crashes and 391 after-installation crashes at the selected sites. For approxi- mately 68 percent of the sites, no crashes were reported for the 6 years of data analyzed. A log-linear model was fit to the data stratified by the year of installation, daytime-versus- nighttime crashes, and before-versus-after installation time period. Overall, nighttime crashes were estimated to have been reduced by 22 percent compared with daytime crashes at the same sites. A disaggregate analysis by year of installation revealed that sites modified in 1976 and 1977 had reductions of 33 percent and 32 percent, respectively. However, sites modified in 1978 showed a 53-percent increase in nighttime crashes, an effect that could not be explained. Perhaps sites most worthy of PRPM installation, and therefore most likely to yield safety benefits, were treated earlier in the program. Single-vehicle crashes were estimated to have been reduced by 12 percent more than other nighttime crash types were. These reduc- tions were found to be independent of ADT and curvature, although it should be remembered that all curves had at least 6 degrees of curvature. Kugle et al. (11) collected 2 years of before data and 2 years of after data at 469 Texas locations varying in length from 0.2 to 24.5 miles (0.32 to 39.4 km). PRPMs were installed between 1977 and 1979. Sixty-five percent of the locations were on two-lane roads; 33 percent on four-lane roads; and the remaining on three-, five-, or six-lane roads. Seventeen sites were subsequently omitted from analysis because they were resurfaced after PRPM installation, which would likely have influenced crash risk. Crashes were sub- classified for analysis by wet weather/dry weather, daytime/ nighttime, and fatal/injury/property-damage-only (PDO). Comparison of wet-versus-dry crashes excluded conditions such as muddy and snowy, but these crashes were included in the total nighttime-versus-daytime analysis. Daytime crashes included daytime, dawn, and dusk crashes, while nighttime crashes included crashes with and without street lighting pres- ent. Crash types potentially affected by PRPMs—namely head-on, sideswipe, and run-off-road—were identified for a separate analysis. In addition, ADT and the number of wet weather days were recorded for each location during the analysis periods. Three evaluation methods were used and are described below. TABLE 2-2 PRPM replacement cycle for the state of Indiana Number of lanes ADT (veh/day) Replacement cycle (years) Two Fewer than 5,000 4 5,000 to 15,000 3 More than 15,000 2 Four or more Fewer than 10,000 4 10,000 to 30,000 3 30,000 to 75,000 2 More than 75,000* 2 * These roadways should be inspected at least once each year. TABLE 2-3 When to schedule PRPM system maintenance for the state of Texas (based on nighttime inspection) For markers spaced at… Maintenance should be scheduled as soon as possible if… 80 ft (24 m) Fewer than two markers are visible 40 ft (12 m) Three or fewer markers are visible TABLE 2-4 Suggested replacement cycles for PRPMs for the state of Texas ADT (veh/day) Replacement cycle (years) More than 50,000 1 More than or equal to 10,000 2–3 Fewer than 10,000 3–4

The first evaluation method involves calculating the cross- product ratio as an overall measure of effectiveness. This method aggregates data from all sites and does not consider noncrash factors, such as ADT. The cross-product ratio mea- sures the relative change in the crash type of interest com- pared with a control group of crashes believed not to be affected by the measure of interest. The control group can therefore be used to control for factors such as changes in ADT and other changes over time that affect crash risk. Night- time crashes were compared with daytime crashes and wet weather crashes compared with dry weather crashes. Using nighttime-versus-daytime crashes for illustration, the cross- product ratio, T, is calculated as (2-1) Where x11 = Crashes in the after period during nighttime, x12 = Crashes in the before period during nighttime, x21 = Crashes in the after period during daylight, and x22 = Crashes in the before period during daylight. The change in crash frequency due to treatment is esti- mated as Percent Change = 100(T − 1) (2-2) The second method, called Gart’s procedure, calculates the cross-product ratio at each individual location and weights each estimate by the total number of crashes at each site for a weighted average estimate of treatment effectiveness. This allows the higher-crash locations to exert more influence on the estimated effectiveness. The third method uses logistic regression, which can include the influence of factors other than the PRPMs in the estimation of effectiveness. The probability of an individual location experiencing a nighttime crash is modeled as a func- tion of time (before or after installation), ADT, and number of lanes. This procedure provides an estimate of effective- ness adjusted for site differences in ADT and number of lanes. Kugle et al.’s analysis of wet weather crashes also included the number of wet weather days as a variable in the model (11). The results of the three methods provided different numeric results, but all methods indicated the same trend for all crash types. The cross-product analysis indicated a 15-percent increase in nighttime crashes and a nonsignificant 1.4-percent decrease in wet weather crashes. Gart’s procedure indicated a 31-percent increase in nighttime crashes and a nonsignifi- cant 1-percent decrease in wet weather crashes. Logistic mod- eling also indicated a significant increase in nighttime and a nonsignificant decrease in wet weather crashes. These effects were found to be consistent for all crash and severity types with the exception of wet weather sideswipe crashes, which T x x x x = 11 22 12 21 8 showed a nonsignificant increase. The authors noted that roughly half of the sites showed a reduction in both nighttime and wet weather crashes, but roughly 10 percent of the sites showed very large increases in total crashes, which may have unfairly skewed the overall results. Mak et al. (14) conducted a study using the same Texas locations as Kugle et al. (11) to reevaluate the safety effect of PRPMs on nighttime crashes. This study screened the original database of 469 locations and eliminated those that underwent major modifications other than the PRPM instal- lation during the evaluation period, modifications that may have influenced the previous study results. Several other locations were not included in the new analysis because they experienced no crashes in either the 2-year before period or the 2-year after period. After screening for these criteria, only 87 of the original 469 locations remained for further analysis. The new analysis focused on individual locations. The day- time crashes were again used as a comparison group to account for any factors that may have influenced crash frequency between the before and after periods but that were not related to the PRPM installation. However, daytime crashes did not include crashes occurring during dusk or dawn; dusk and dawn crashes that were eliminated from the analysis were reported to be about 1–3 percent of the total crashes. A statistical pro- cedure, based on the cross-product ratio, was used to measure the effect of PRPMs at individual locations, Z. This procedure is based on a test statistic: (2-3) Where x11 = Crashes in the after period during nighttime, x12 = Crashes in the before period during nighttime, x21 = Crashes in the after period during daylight, and x22 = Crashes in the before period during daylight. Z was calculated for each location. If there are no safety effects, Z will be normally distributed with a mean of 0 and a variance of 1. A positive value of Z would indicate an increase in nighttime crashes relative to daytime crashes, a negative value indicates a relative decrease, and a value of zero indicates no change. Of the 87 locations, 56 (64.4 percent) showed a relative increase in nighttime crashes, 30 (34.5 per- cent) showed a relative decrease, and 1 (1.1 percent) showed no change. Using a confidence level of 10 percent to check for significance, 4 locations (4.6 percent) showed signifi- cant reductions in nighttime crashes relative to daytime crashes, 9 (10.3 percent) showed significant increases, and 74 (85.1 percent) showed nonsignificant changes in night- time crashes relative to daytime crashes. T x x x x = 11 22 12 21 , Z T x x x x = + + + ln( ) 1 1 1 111 12 21 22

The effect of PRPMs on crash severity was also studied at 37 locations that had a minimum of 30 crashes in the before or after period or in the two periods combined. Two severity indexes were calculated for each site separately for nighttime and daytime crashes: • Severe = (percent fatal or incapacitating-injury crashes in after period) − (percent fatal or incapacitating-injury crashes in before period). • Injury = (percent fatal, incapacitating-injury, or non- incapacitating-injury crashes in after period) − (percent fatal, incapacitating-injury, or non-incapacitating-injury crashes in before period). A logit model was then used to test for statistically signif- icant differences using the daytime crashes as a comparison group. None of the 37 sites showed a significant change in the percentage of severe crashes, perhaps due to low numbers of these crashes. For injury crashes, 4 locations showed a sig- nificant decrease in nighttime crash severity, 1 showed a sig- nificant increase, and 32 showed no significant change. Locations that showed a significant increase (nine total) or decrease (four total) in crash frequency were further exam- ined in an attempt to identify crash characteristics that might be associated with PRPMs. Not enough data were available for statistical tests, but an examination of the relative pro- portions between the before and after periods indicated that for the sites showing a significant increase in crashes, the pro- portion of nighttime multivehicle crashes increased and the proportion of nighttime fixed-object crashes decreased. For the locations showing a significant decrease in crashes, this same increase in multivehicle crashes and decrease in fixed- object crashes was found for the daylight hours but not for nighttime hours. For both groups, the proportion of nighttime crashes occurring on horizontal curves greater than 2 degrees increased. A number of roadway characteristics were also examined for their effect on the influence of PRPMs using the same groups of nine (significant increase) and four sites (significant decrease), but no strong evidence that any of the variables interacted with PRPM installation was found. The examined characteristics included the following: • Intersection type (none, interchange, T-intersection, four- leg intersection, or multiple intersection), • Whether the roadway was within the city or outside the city, • Horizontal curvature (less than 1 degree, 1 to 3 degrees, or more than 3 degrees), • Grade (less than 3 percent or more than 3 percent), • Structures (none, culvert, or bridge), • Number of lanes (less than or equal to four, or more than four), and • Whether the roadway was divided or undivided. 9 Griffin (15) analyzed the same data as Mak et al. (14), which is a subset of the Texas data originally used by Kugle et al. (11) using a different statistical approach. Griffin quantified the safety effect of PRPMs on nighttime crashes at 86 locations using daytime crashes as a control group. One of the loca- tions used in the previous analysis was not included in this study because it could not be located. The overall, or aver- age, effect of PRPM installation on nighttime crashes was estimated by calculating a weighted log odds ratio. The log odds ratio, L, is calculated by taking the natural logarithm of T, defined previously in Equation 2-1. The weight for the log odds ratio at each site, w, is calcu- lated as (2-4) Where x11 = Crashes in the after period during nighttime, x12 = Crashes in the before period during nighttime, x21 = Crashes in the after period during daylight, and x22 = Crashes in the before period during daylight. The weighted log odds ratio, Lavg, is thereafter calculated as (2-5) Where L = ln(T) The average effect is equal to the antilogarithm of Lavg, and the standard error of L, Lse, is equal to (2-6) Using this methodology, the expected change in nighttime crashes following the installation of PRPMs was estimated to be a 16.8-percent increase, with the 95-percent confidence limits between a 6.4- and 28.3-percent increase. Pendleton (16) used both “classical” and empirical Bayes before-and-after methods for evaluating the effect of PRPM nighttime crashes on undivided and divided arterials in Michi- gan. Seventeen locations totaling 56 miles (90 km) served as installation sites, and 42 sites totaling 146 miles (235 km) were used as control sites where PRPMs were not installed. Crash data for 2 years prior to installation and 2 years after installation were used for two categories of analysis. The first category used as a control group daytime crashes at the instal- lation sites, which were assumed to be unaffected by the instal- lation of PRPMs. Daytime crashes did not include crashes that occurred 1 hour before and 1 hour after both sunrise and L w se = ∑ 1 L wL w avg = ∑ ∑ w x x x x = + + +( ) 1 1 1 1 111 12 21 22

sunset, a total of 4 hours per day. The second category used nighttime crashes at control sites as a control group. Pendleton made the following conclusions: • Undivided roadways showed an increase in nighttime crashes and divided roadways showed a decrease in night- time crashes when analyzed separately. Whether a high- way was divided was concluded to be the most significant road characteristic affecting the effectiveness of PRPMs. • Using daytime crashes at treated sites as a comparison group yielded larger reductions (or smaller increases) in crashes than when nighttime crashes at untreated sites were used as a comparison group. The issue of which comparison group to use stayed unresolved. • The empirical Bayes methodology generally produced smaller reductions (or larger increases) than the simple or “classical” before-and-after methodology. This con- clusion usually is an indication that regression-to-the- mean was at play and accounted for by the empirical Bayes methodology. • Exposure should be properly accounted for, and the researchers lamented the fact that the estimates of night- time traffic volume were only approximations. The study also revealed the difficulties of using crash rates (crashes per unit of exposure) to control for exposure differences. These difficulties arise from the nonlinear relationship between crashes and exposure that indicates that these rates can change because of volume changes and not necessarily because of a treatment. New York State DOT (17) undertook a safety assessment of PRPMs in New York to review the DOT’s policy on PRPM installation. The DOT used a simple before-and-after study design in which numbers of crashes before and after treat- ment were compared without controlling for other factors. Two analyses were undertaken using this simple before-and- after design. The first analysis, at 20 sites, targeted PRPMs at sections of unlit suburban and rural roadways with pro- portionately high numbers of nighttime and nighttime wet weather crashes. Overall, there was a nonsignificant decrease of 7 percent for total crashes, a highly significant decrease of 26 percent for nighttime crashes, and a significant decrease of 33 percent for nighttime wet weather crashes. Furthermore, there was a significant reduction of 23 percent in all guidance- related crashes, which are crashes resulting from a vehicle leaving its assigned travelway (e.g., run-off-road, head-on, encroachment, and sideswipe). There was also a 39-percent reduction in nighttime guidance-related crashes. The second analysis looked at PRPMs installed non- selectively over 50 long sections of highway. The analysis revealed that nighttime crashes were reduced by a nonsignifi- cant 8.6 percent, that total crashes were reduced by a statis- tically significant 7.4 percent, and that nighttime wet weather crashes increased by a nonsignificant 7.4 percent. Thus, New York State DOT recommended that PRPMs be installed selec- tively “when their use is likely to reduce crash frequency cost 10 effectively by improving delineation during nighttime wet weather conditions.” It further stated that PRPMs should be installed only at locations having high frequencies of wet weather, nighttime, guidance-related crashes. Orth-Rodgers and Associates, Inc. (18), used the same “odds ratio” methodology as Griffin (15) to evaluate the effects of both raised and recessed pavement markers on nighttime crashes at 91 Interstate highway locations in Penn- sylvania. PRPMs were installed at these sites between 1992 and 1995, and crash data from 1991 to 1996 were used in the analysis. Daytime crashes at the same sites were used as a comparison group. Sites that had no crashes in any of the day- time or nighttime periods before or after PRPM installation were eliminated since a zero value would render the odds ratio meaningless. This omission creates a subtle bias toward underestimation of effects if the realization of zero crashes at a site in the after period is due to PRPM installation. This underestimation is exaggerated by the fact that the after peri- ods were, on average, much shorter than the before periods and were therefore more likely to contain zero crashes. Sites in urban and lit areas were also eliminated, assumed by the authors as “not good candidates for an analysis of this type.” Several crash types were excluded because they were con- sidered to be unrelated to PRPMs (e.g., crashes that hap- pened during dusk, dawn, or unknown lighting condition; crashes that occurred in weather conditions other than rain or “no adverse conditions”; crashes that occurred when the road surface condition was other than dry or wet; and crashes for which the impact type was “unknown”). Results indicated a 12.3-percent increase in nighttime crashes (95-percent confidence limits of 1.1 and 24.8 per- cent) for all sites, a nonsignificant 1.2-percent decrease for locations with raised pavement markers, and a significant 20.1-percent increase (95-percent confidence limits of 5.5 and 36.9 percent) for locations with recessed pavement mark- ers. The authors suspected that the small decrease in night- time crashes due to raised PRPMs might have been because there was a positive effect (i.e., a reduction in crash fre- quency) on the daytime crashes that was used for the com- parison group. Additional results were obtained for two crash subsets. Nighttime wet condition crashes also showed large increases from 30 to 47 percent (confidence limits not reported), depend- ing on the comparison group of crashes used (daytime wet condition, nighttime other, or all daytime crashes). Nighttime wet road sideswipe and fixed-object crashes were estimated to have increased 56.2 percent (confidence limits not reported) using nighttime dry road sideswipe and fixed-object crashes as a comparison group. Not much emphasis was placed on these additional results since these increases could be exaggerated by a positive effect of PRPMs on the comparison sites. Table 2-5 summarizes the review of seven relevant evalu- ations of the safety effects of PRPMs, measured in terms of reductions or increases in crashes (two of the seven studies are re-analyses of subsets of data previously analyzed). All but one of the studies listed in Table 2-5 used daytime crashes

11 TABLE 2-5 Summary of literature on the safety effectiveness of PRPMs Study Ref. /Location Site Type Installation Location I – Installation Period B – Before- Period Length A – After- Period Length Sample Sizes for Treatment and Comparison Groups Dependent Variable Independent Variables Analyzed Comparison Group Other Notes Estimated Effects Wright et al. 1982 (13) Georgia Horizontal curves on two-lane highways in excess of 6 degrees of curvature Centerline I – 1976-1978 B – 1 to 3 years A – 2 to 4 years Treatment – 662 locations Comparison – same as treatment group Total nighttime crashes ADT, degree of curvature Total daytime crashes Both raised and recessed reflective markers were used; at some locations warning signs, chevron markers or other guidance devices were installed 22% reduction in nighttime crashes; single-vehicle crashes reduced 12% more than other nighttime crashes; reductions independent of ADT or horizontal curvature for curves with degree of curve greater than 6 Kugle et al. 1984 (11) Texas Two- , three-, four-, five -, and six- lane roadways Does not specify I – 1977-1979 B – 2 years A – 2 years Treatment – 452 locations Comparison – same as treatment group Total nighttime crashes, some analysis by crash and severity ADT, number of lanes, number of wet weather days Total daytime crashes None 15 to 31% increase in nighttime crashes; no significant effect on wet weather crashes Mak et al. 1987 (14) Texas Two- , three-, four-, five -, Does not specify I – 1977-1979 B – 2 years Treatment – 87 locations Comparison – Total nighttime crashes, Intersection type, within/outside city, horizontal curvature, grade, structures, number of lanes, divided/ undivided Total daytime crashes Used a subset of the data from Kugle 4.6% of locations wed significant reductions, 10.3% showed significant increases, 85.1% showed nonsignificant effects and six- lane roadways A – 2 years same as treatment group some analysis by crash and severity types et al., 1984 (11) Griffin, 1990 (15) Texas Two- , three-, four-, five -, and six- lane roadways Does not specify I – 1977-1979 B – 2 years A – 2 years Treatment – 86 locations Comparison – same as treatment group Total nighttime crashes None Total daytime crashes Used a subset of the data from Kugle et al., 1984 (11) 16.8% increase in nighttime crashes, with the 95% confidence interval between a 6.4 and 28.3% increase. Pendleton, 1996 (16) Michigan Divided and undivided arterials Centerline on undivided arterials, lane lines on divided arterials I – 1989 B – 2 years A – 2 years Treatment – 17 locations totaling 56.11 mi (90.3 km) Comparison – 42 locations totaling 146.28 mi (235 km) Total nighttime crashes Divided/ undivided and VMT (vehicle miles traveled) used in empirical Bayes analysis Total daytime crashes, total nighttime crashes at comparison sites None No significant effect, direction of effect positive or negative dependent on method used and access control New York State DOT, 1989, 1997 (17, 19) Suburban and rural roadways Does not specify I – unknown B – unknown A – unknown Selective Installation: Treatment – 20 locations totaling 26 mi (41.84 km) Comparison – none used Nonselective Installation: Treatment – 50 locations Comparison – Total crashes, total nighttime crashes None None Regression to the mean is cited as being a factor 26% decrease in nighttime crashes when placed selectively, no significant effect when installed nonselectively none used Orth-Rodgers and Associates, Inc., 1998 (18) Pennsylvania Interstate highways in rural non- illuminated areas Does not specify I – 1992-1995 B – 1-3 years A – 1-3 years Treatment – 33- 76 locations depending on crash type studied Comparison – same as treatment group Total nighttime crashes, nighttime wet road, nighttime wet road sideswipe fixed-object None Total daytime crashes, daytime wet road, daytime wet road sideswipe or fixed-object Both raised and recessed reflective markers were used 18.1% overall increase in nighttime crashes, nighttime wet condition crashes increased from 30 to 47%, nighttime wet road sideswipe or fixed-object increased by 56.2% New York

as a comparison group for nighttime crashes, based on the assumption that only nighttime crashes would be affected by PRPMs. As discussed in Section 2.3, there is evidence that PRPMs affect driver behavior during daytime as well, man- ifested by changes in positioning in the lane and significant reductions in lane encroachments, which would be expected to impact both head-on and run-off-road crashes. Conse- quently, the use of daytime crashes as a comparison group is inappropriate. Table 2-5 shows both significant reductions and increases in crash frequency. Indeed, the two largest studies show opposing effects—one with 662 treatment locations (13) showing a 22-percent reduction in nighttime crashes, and the other with 452 treatment locations (11) showing a 15- to 31-percent increase in nighttime crashes. Re-analysis (15, 14) of the second study, with its troubling result, continued to show a statistically significant increase in nighttime crashes at some locations. As will be seen in Section 2.3, there are mixed findings with respect to speed and an indication that speed effects may be site specific. Changes in speed, along with the effects of PRPMs on daytime encroachments, may be factors in the mixed safety effects. 2.2.2 Methodological Problems in Past Research The relative safety at any location is a function of all road- way, environmental, and driver characteristics. A change in any of these factors from the before to after period affects safety. In order to derive an accurate estimate of the safety effect of PRPM installations, it is important to separate the effect of other changes, including the changes described in the following sections. 2.2.2.1 Changes in Traffic Volumes Safety directly relates to traffic volumes. As a result, the dif- ference in traffic volumes between the before and after peri- ods affects the expected difference in the number of crashes between the before and after periods. In most of the previ- ous studies reviewed, traffic volumes were not accounted for explicitly. Daytime crashes have most often been used to control for changes in safety, on the assumption that these are unaffected by the PRPM installation. In the treatment- comparison experimental design used by several researchers, it was assumed that traffic volume changes are controlled for because the percentages of AADT during day and night should not change significantly in the before and after time periods (18). This may be a reasonable way of accounting for traffic volume changes if this assumption is met, providing that the changes are small and that the relationship between crashes and traffic volume is approximately linear. In the studies reviewed, it was unclear if these provisions were in fact met. It seems reasonable that one should not rely on such assump- tions and that one should seek explicit ways of accounting for 12 traffic volume changes in both the treatment and comparison groups since such volumes may be relatively easy to acquire and influential in the evaluation results. 2.2.2.2 Time Trends Areawide safety changes over time because of many fac- tors, such as weather conditions, driver demographics, and vehicle technology. Reporting levels also directly affect crash data. Often either the minimum damage dollar value changes for PDO crashes or the reporting level by police changes. PDO crash data from jurisdictions that have switched from 100-percent police reporting to self-reporting crash data must be used carefully when accounting for the safety effect of PRPMs. Again, most of the previous studies attempted to account for the safety effect of PRPMs by using daytime crashes at the same sites as a comparison group. However, if the installations of PRPMs have a safety effect on daytime crashes and/or the time trends between daytime and night- time crashes differ, using daytime crashes at the same sites as a comparison group will result in errors in the estimate of safety effectiveness. 2.2.2.3 Regression-to-the-Mean If PRPMs were installed at a location experiencing a ran- domly high number of crashes in the before period, then the number of crashes in the after period would be expected to decrease with or without the installation of PRPMs. This phenomenon (known as regression-to-the-mean, or RTM) is often a factor when study sites are selected based on crash history. Not only could RTM exist for the crash type or loca- tions of interest, but it could also exist in the comparison group, and this existence could exaggerate the positive effects of a measure. For example, Orth-Rogers and Associates (18) cite a study by Khan (20) in which 184 sites were selected from high-hazard locations having four or more crashes in 1 year before the installation of PRPMs. At a group of con- trol locations where PRPMs were not installed, it was found that the total number of crashes increased. However, at the treated sites, both nighttime and daytime crashes were reduced. It is clear that, given the site selection criterion, RTM will exaggerate the positive effects noted in the Khan study (20) and may even explain in entirety the reduction in daytime crashes. However, the increases in the control group may also be due to RTM because these locations may have been untreated because they fell into a group that had fewer than the average number of crashes in 1 year. This RTM would exaggerate the effects of PRPMs even more. Only one of the previous studies, Pendleton (16), directly accounted for RTM effects. The treatment-comparison exper- imental design can, in principle, use comparison sites to con- trol for RTM, but the treatment and comparison sites need to be matched on the number of crashes. In practice, controlling

for RTM using a treatment-comparison experimental design is achievable only if sites are randomly assigned to a treat- ment and comparison group—a desideratum that is difficult to accomplish in road improvement programming. Alterna- tively, there will be no RTM concern if the fact that a site had a higher than usual crash history is not used in the selection of the sites for treatment. For example, in the “before and after design with yoked comparison” used by Griffin (15) and Orth-Rogers and Associates (18), RTM bias is thought to be eliminated by not using the number of crashes as a criterion for selecting sites. It is unclear how RTM bias was eliminated. This strategy, however, is not realistic because it defeats the purpose of safety improvement programs since measures are likely to have the largest safety benefits where a safety con- cern is manifested in a high crash frequency. 2.2.2.4 Other Measures Simultaneously Applied Zador et al. (21) acknowledged the difficulty of identify- ing the effect of one treatment when multiple treatments have been applied. This difficulty presents the following method- ological challenge: to discard data where changes in addition to PRPM installation occurred during the study period or to try to isolate the effects due to PRPMs. For the latter option, a promising methodology recently applied by Feber et al. (22) could be considered. 2.2.2.5 Selection of the Comparison Group— the Problem of Spillover and Migration Effects Treatment-comparison experimental designs are commonly used to control for effects not due to the treatment. The treat- ment effects would be underestimated if, as some of these studies have found, there were a decrease in target crashes at comparison sites that was due to spillover effects of the treat- ment. Measures such as red light cameras are believed to have such effects. The importance of this point is emphasized by Orth-Rogers and Associates (18) in their analysis of the Pennsylvania data. As indicated earlier, the authors suspected that PRPMs may have had a positive effect on the daytime crashes used for the comparison group that generated the result that PRPMs caused only a marginal decrease in nighttime crashes. The authors further concluded that if this impact on the compari- son group were true, then the fundamental basis of the analy- sis conducted by Griffin (15) as well as on their own study is questionable. In contrast to the underestimation caused by spillover, treat- ment effects would be overestimated if there were crash “migration” (i.e., an increase in crashes at the comparison sites due to the compensatory behavior of drivers). The installation of all-way stop control and other speed-control measures are believed to sometimes cause vehicles, and therefore crashes, to 13 “migrate” to other sites; thus, it is conceivable that sites adja- cent to PRPM installation sites will experience such migra- tion effects. 2.3 LITERATURE REVIEW OF HUMAN FACTORS ISSUES AND PRPMS The following subsections review the human factors issues related to the use of PRPMs: • Driver needs with respect to delineation and visibility, • Visibility of PRPMs, and • Driver behavior in response to PRPMs. 2.3.1 Driver Needs with Respect to Delineation and Visibility Pavement markings and delineation devices provide an important guidance function for drivers, especially at night. Pavement markings and delineation devices provide drivers with information about the vehicle position within the lane and information about which lanes are available for use. Pavement markings and delineation devices also provide the driver with a preview of upcoming changes in the roadway geometry, including curves, lane drops, narrowing, the start and end of passing zones, crosswalks, and intersections. There is a perception-reaction time delay between seeing a change in the road path and responding to it and between making a steering input and the vehicle responding; there- fore, several seconds of preview are required for good lane positioning. Good delineation generally results in better driver performance and greater driver comfort. Driver requirements for delineation have been established through studies of lane tracking given various driver preview distances and through studies involving the recording of driver eye movements. Driver preview distance may be modified by blocking parts of the forward view through the windshield or by simulating reduced visibility conditions in a driving simulator. In actual vehicles on a tangent section of road, McLean and Hoffman (23) found that, at 31 mph (50 km/h), improving sight distance beyond 2 seconds did not further improve lane position control. On a highway, at speeds of 50 to 68 mph (80 to 110 km/h), eye movement recorders showed that drivers looked about 3 seconds ahead of the vehicle (24). According to a Commission Internationale de l’Eclairage (International Commission on Illumination, or CIE) report on visual aspects of road markings (25), Farber et al. (26) found that a minimum of 5 seconds’ preview time was necessary to allow for efficient, anticipatory steering behavior. Based on these and other studies, the CIE report (25) rec- ommends a minimum practical preview time of 3 seconds and a desirable preview time of 5 seconds. The sharper the curve, the greater the preview distance required to allow for

the time it takes to perceive and react to the curvature by dropping speed. Surface markings are recognized as sufficient for providing 2 to 3 seconds of preview time, while longer preview times require the use of PRPMs or post-mounted delineators (2). At 60 mph (96 km/h), a preview time of 2 to 3 seconds would be equivalent to a driver being able to see 2 to 3 PRPMs ahead at the recommended spacing for tangents (every 80 ft or 24 m), and 4 to 6 PRPMs ahead at the recommended spac- ing for curves (every 40 ft or 12 m). 2.3.2 Visibility of PRPMs The visibility of PRPMs depends on aspects of the device, its placement, the vehicle head lighting, the highway geom- etry, and the driver visual capabilities. Drivers detect the presence of a delineator by means of slight differences in brightness between the delineator and the road surface. This difference or contrast, C, is defined as (2-7) Where LT = target luminance and LB = background luminance. Once contrast reaches a certain level, known as the thresh- old contrast, it is just detectable to the viewer. During the day, the visibility of delineators depends only on the contrast between the delineator and the pavement background. At night, visibility depends on the light from headlights as well as on the retroreflectivity of the delineator. Retroreflection means that the light is reflected back at the same angle at which it is projected. If light from the headlights were to be perfectly retroreflected, it would not reach the driver’s eyes, which are above the headlights. Since retroreflection is imper- fect, some of the light reaches the driver’s eyes, increasing the contrast between the delineator and the low-reflectance (pave- ment) background. The higher the percentage of light that is reflected back to the driver’s eye, the greater the contrast and the further away the delineator will be seen. 2.3.2.1 Device Features Device design and condition both have strong effects on visibility distance. With respect to device design, Blaauw and Padmos (27) compared three types of PRPMs that varied in the arrangement and number of lenses: • Metallic mounting with three large, biconvex lenses (Category A), • Plastic mounting with 21 small, biconvex lenses (Cate- gory B), and C L L L T B B = − 14 • Plastic mounting with corner-cube lenses (Category C). Visibility distances were determined through measure- ments of the optical characteristics of the PRPMs, combined with data from experiments with subjects. Figure 2-1 shows visibility distances for various types of delineators, including the three types above: under different atmospheric visibility distances, including clear weather (Z = 9.3 miles or 15 km), moderate fog (Z = 0.62 miles or 1 km), and heavy fog (Z = 0.12 miles or 0.2 km); for low-beam headlamps in both new and used condition; and under wet and dry pavement conditions. The line denoted “V85 = 100 km/h” (62 mph) indicates the distance necessary to provide 5 seconds of preview time for Figure 2-1. Predicted visibility distances immediately after application for all markings on a dry and wet pavement (Z = atmospheric visibility distance) (27). 250 200 150 100 50 0 Z = 0.2 km V85 = 80 km/h V85 = 70 km/h new in practice dipped headlamps paint thermo- plast profile 1 profile 2 A raised B pavement C markers 250 200 150 100 50 0 Z = 1 km V85 = 100 km/h V85 = 80 km/h vi si bi lity d ist an ce (m ) 300 250 200 150 100 50 0 Z = 15 km V85 = 100 km/h V85 = 80 km/h dry wet NEW

the 85th percentile of the velocity distribution on a rural road with a 62-mph (100-km/h) speed limit. As can be seen in Fig- ure 2-1, the PRPMs with corner-cube reflectors (Category C) had visibility superior to the other two categories (A and B). Headlights deteriorate over time, causing visibility distances to shorten. This phenomenon can be seen in Figure 2-1, where values that are designated “new” have a calculated visibility based on isocandela diagrams provided by the headlamp man- ufacturer and values that are designated “in practice” were used in the experiment. The distances are based on the mea- sured retroreflection coefficients. Requirements for minimum visibility distances are given for rural roads with V85 veloci- ties of 50 and 62 mph (80 and 100 km/h). The requirements for heavy fog (Z = 0.12 miles, or 0.2 km) are lower because of the lower effective speeds on roads that experience foggy conditions (27). While PRPMs provide better visibility than painted or tape lane markings, PRPMs deteriorate more rapidly over time. Figure 2-2 shows visibility distances 22 months after appli- cation on a highway lane with an AADT of 3,062 veh/day. As can be seen, visibility distances for in-service devices are reduced by as much as half of that of newly installed devices. However, even under dry conditions, the visibility of PRPMs is still better than the visibility of paint. 2.3.2.2 Environmental Conditions As shown in Figures 2-1 and 2-2, rain and atmosphere trans- mittance strongly affect delineation visibility. The atmospheric conditions examined include clear weather, moderate fog, and heavy fog. Reductions in visibility of PRPMs due to rain were on the order of 10 to 20 percent depending on the type of marker and the environmental conditions. Reductions due to decreased atmosphere transmittance were larger, ranging from 40 to 60 percent. However, the visibility of the PRPMs was still better than the visibility of paint under all conditions, even after the PRPMs had been in service for 22 months. 2.3.2.3 Headlighting Headlight patterns affect delineator visibility. A typical low-beam pattern is shown in Figure 2-3. Headlights are aimed to the right and down a few degrees to avoid glare for oncoming drivers. This means that more light falls on the right side of the road than on the left side, and, with low- beam headlights, delineators on the right will be visible at longer distances than those on the left. Headlights deteriorate over time, causing visibility distances to shorten, as can be seen in Figures 2-1 and 2-2. 2.3.2.4 Road Geometry Road geometry affects delineator visibility. The more the face of the delineator is aligned perpendicular to the line of 15 sight of the driver, the more visible the device will be. On curves, maximum visibility will be obtained when the PRPM face is aligned perpendicular to the tangent of the curve. Recommended spacings between PRPMs on tangents and curves are in the “Roadway Delineation Practices Hand- book” (2). 2.3.2.5 Driver Characteristics Contrast sensitivity. Driver characteristics, mainly con- trast sensitivity, affect delineator visibility. Sensitivity to con- trast varies greatly among drivers, even among drivers with “normal” acuity of 20/20. Because of differences in contrast Figure 2-2. Visibility distance 22 months after application of the markings (27). 250 200 150 100 50 0 Z = 0.2 km V85 = 80 km/h V85 = 70 km/h new in practice dipped headlamps paint thermo- plast profile 1 profile 2 A raised B pavement C markers 250 200 150 100 50 0 Z = 1 km V85 = 100 km/h V85 = 80 km/h vi si bi lity d ist an ce (m ) 300 250 200 150 100 50 0 Z = 15 km V85 = 100 km/h V85 = 80 km/h dry wet 2.106 vehicles

sensitivity, driver detection distances for delineation devices can vary by a factor of 5 to 1. As drivers age, contrast sensi- tivity declines, reducing preview distances available and lead- ing many older drivers to reduce nighttime driving. Age. In an FHWA study directed at the needs of older drivers with respect to delineation of horizontal curves, 16 Pietrucha et al. (28) examined the response of drivers in three age groups (18–45, 65–74, and 75+) to 25 delineation treatment combinations. Table 2-6 describes 12 treatments. The baseline treatment (Treatment 1) was a 4-in. (100-mm) yellow center- line with a measured coefficient of retroreflected luminance (RL) of 100 mcd/m2/lux (referred to as an in-service brightness level). Left and right curves were studied with a radius of Figure 2-3. Isocandela diagram of a typical U.S. low-beam pattern superimposed on a road scene (18). Treatment Number Centerline Treatment Edgeline Treatment Off-Road Edge Treatment 1 4-in. Yellow Line None None 2 4-in. Yellow Line 4-in. Structured Line None 3 4-in. Yellow Line + Yellow PRPMs None None 4 4-in. Yellow Line + Yellow PRPMs White PRPMs None 5 4-in. Yellow Line None Normal Mount Chevrons 6 4-in. Yellow Line 4-in. White Normal Mount Chevrons 7 4-in. Yellow Line None Standard Flat Posts (Hi-Intensity) 8 4-in. Yellow Line 4-in. White Standard Flat Posts (Hi-Intensity) 9 4-in. Yellow Line None Full Reflection Posts (Hi-Intensity) 10 4-in. Yellow Line None T-Posts (Hi-Intensity) 11 4-in. Yellow Line + Yellow PRPMs None T-Posts (Hi-Intensity) 12 4-in. Yellow Line 4-in. White T-Posts (Engineering) TABLE 2-6 Details of 12 delineation treatments (28)

500 ft (152 m). Treatments on curves studied included PRPMs on the centerline (Treatments 3 and 11) or on the edgeline and centerline simultaneously (Treatment 4) as well as treat- ments with and without chevrons and post-mounted delin- eators. Measures were recognition distance and time spent looking at the roadway. Measurements were taken using a visual occlusion device and using subject assessment. The first phase of the study involved a laboratory test of simulated nighttime driving. In the second phase of the study, a subset of the best treatments was then field-tested with the youngest and oldest age groups (see Figure 2-4). As is so often the case, the treatment that improved performance for the older drivers also improved performance for younger drivers. The treatment with the highest recognition distance for both groups was Treatment 12, which did not use PRPMs. Treatment 12 had a 4-in. (100-mm) yellow centerline with a measured coefficient of retroreflected luminance (RL) of 100 mcd/m2/lux (in-service level of brightness), a 4-in. (100-mm) white edgeline, and T-posts with engineering-grade reflec- tivity with standard spacing (65 ft or 19.8 m). There were significant differences between left- and right- curve recognition distances for some treatments and between older and younger drivers for other treatments. On average, the older drivers had 14 percent less recognition distance than the younger drivers. Furthermore, it has been demonstrated that older drivers who volunteer for testing are likely to have sub- stantially better vision than the average older driver. Conse- quently, differences amongst the driving public between older and younger drivers are likely to be much more pronounced. 2.3.2.6 Visibility Distance and Benefit-Cost Analysis In the Pietrucha et al. study (28), a benefit-cost analysis that used visibility distance to determine the benefit, examined the 17 best of 25 initial delineation treatments. The visibility distances for six of the best treatments (Treatments 5, 6, 9, 10, 11, and 12) did not differ with respect to statistical significance. Amongst those treatments, the one recommended on a benefit-cost basis was the treatment with a 4-in. (100-mm) yellow line, no edge- line, and high-intensity post-mounted delineators (Treatment 10). The treatments with PRPMs were more costly for an equal visibility benefit and therefore were not recommended. 2.3.3 Driver Behavior in Response to PRPMs In addition to measures of visibility distance, various mea- sures of driver behavior have been used to evaluate the effec- tiveness of PRPMs. These measures include visual workload (as determined by number of looks required for comfortable driving), speed, speed variation, lane position, lane position variation, and encroachments into adjoining lanes. Some stud- ies involve measuring the behavior of a group of subjects on test courses, while others involve measuring the behavior of unsuspecting drivers on public roads where test installations have been set up. The greatest number of studies of the impact of delineators on driver behavior were directed to horizontal curves. PRPMs on tangents and on ramps have been examined in one study (29). Another study concerned gore areas and deceleration lanes (30), and two studies looked at approaches to narrow bridges (31, 32). These studies are described in the paragraphs below. 2.3.3.1 Driver Lane Position and Speed for PRPMs on Curves Table 2-7 summarizes the effects that PRPMs on curves have on driver behaviors. In a before-and-after study compar- ing delineation with and without centerline snowplowable Old Young 1 0 100 200 300 400 500 600 700 800 900 1000 4 3 2 7 8 TREATMENT M EA N R EC O G N IT IO N D IS TA N C E (ft ) 5 11 6 9 10 12 Figure 2-4. Recognition distance: old versus young groups (combined curves) (28).

PRPMs, Mullowney (30) measured impacts on encroach- ments, on speeds, and on speed variance. Only nighttime data were collected for this study. Speeds were measured using a handheld radar unit. An on-site observer collected encroach- ment data. Centerline encroachments were measured at three sites, including one control site. Two of the three sites were on the same roadway. Edgeline encroachments were mea- sured at two sites on the same roadway, one of which was a control site, presumably the same delineation but without PRPMs, although this presumption was not stated explicitly. After the PRPMs were installed at the treatment sites, they were compared with the control sites, and a statistically sig- nificant reduction (p < 0 .01, p < 0.05) of 12 percent (site with street illumination) and 3.7 percent (site without illumina- 18 tion) in centerline encroachments was measured. There was also a statistically significant reduction (p < 0.01) of 5.7 per- cent in edgeline encroachments at the sites with PRPMs after installation as compared with before. There was no statisti- cally significant difference at two control sites for this com- parison. There is no indication of how long after installation measurements were made, nor is it indicated where installa- tions were centerline only or a combination of edgeline and centerline. The former is assumed. During Mullowney’s study (30), speeds were collected at two sites, both on horizontal curves. At the first site, speeds were collected for traffic in both directions at four different locations around the curve. The presence of PRPMs appears to have resulted in a smoother speed profile through this site TABLE 2-7 Summary of literature on driver performance and PRPMs (horizontal curves) Study Ref. /Location Installation Location PRPM Spacing Effect on Speed Effect on Lane Position Other Mullowney (1982) (30) New Jersey (Night only) Centerline, edgelines separately NA Smoother speed profile through the site Centerline and edgeline encroachments were reduced by significant amounts (3.7 to 12%) When combined with illumination, greater reduction of encroachments was measured Niessner (1984) (31) USA (Day and night) Centerline, edgeline Recommended: • 80 ft (24 m) on curves up to 3 degrees • 40 ft (12 m) on curves up to 15 degrees • 20 ft (6 m) on curves with more than 15 degrees of curvature • No significant difference in mean speeds – nighttime 85th percentile speeds reduced significantly • Speeds increased on one approach and decreased on the other • Smoother speed profile (only night results) • Centerline encroachments reduced by 50% (daytime and nighttime) • Vehicle placement variability reduced significantly – vehicles shifted significantly toward centerline during daytime and toward edgeline at nighttime • Centerline and edgeline encroachments reduced significantly (only nighttime results) Mixture of centerline and edgeline markings appear confusing at some sharp curves Agent & Creasey (1986) (33) Kentucky (Day and night) Centerline • 10 ft (3 m) apart • 20 ft (6 m) apart Daytime and nighttime speeds reduced (p < 0.01) Decreased encroachments for daytime and nighttime (no significance testing) None Zador et al. (1987) (21) 54 sites in Georgia & New Mexico (Day and Both sides of centerline (in conjunction with other measures) • 80 ft (24 m) apart • 40 ft (12 m) apart for sharper curves • Overall mean speed increase of 1 km/h at nighttime • Daytime and nighttime mean speed increase 30 m before and 30 m into curve (from graph) • 6 cm away from centerline: Mean shift 12 cm away from centerline, 30 m before curve • Mean shift 21 cm away from centerline, 30 m into curve (similar effects both daytime None night ) and nighttime) Krammes & Tyer (1991) (34) 5 sites USA (Night only) Centerline NA • Speeds were higher with new PRPMs in place • Speeds went down over time • Vehicles placed further from centerline • Fewer opposite lane encroachments None Hammond & Wegmann (2001) (35) 2 sites in Tennessee (Day only) Both sides of centerline • 40 ft (12 m) apart • 20 ft (6 m) apart No significant difference • Significant reduction in encroachments • No significant difference between 6- and 12-m spacing None

in both directions, as evidenced by less variation in speed between the data collection points after installation of PRPMs than before installation of PRPMs. Speeds were higher at the apex of the curve than at the curve entrance or at the exit of the curve. At the second site, speeds were collected at three locations around the curve. The speeds at the entrance and exit of the curve after PRPM installation were slower than those speeds before PRPM installation, resulting in a smoother speed profile throughout the site. (Note that the speed data were presented only in graph form—no statistical analysis was performed.) The effects of PRPMs at three horizontal curve sites were investigated by Niessner (31) before and after the installation of PRPMs. At the first site, which was an S-curve, conven- tional markers were spaced 40 ft (12 m) apart, two on the cen- terline and one on each edgeline. There was no statistically significant difference in daytime or nighttime speeds, but nighttime 85th percentile speeds were significantly reduced. Centerline encroachments were reduced by 50 percent dur- ing both day and night. At the second site, a single row of PRPMs was installed on the centerline as well as on both edgelines. Speed measurements were taken at three points for both directions. Mean speeds decreased for one approach to the curve, but increased for the other approach. During the daytime, the vehicle lane position shifted significantly toward the center of the curve, while at night, the vehicle lane position shifted toward the edgeline. At the third site, which was an S-curve, snowplowable PRPMs were installed along both edgelines and on both sides of the centerline. Speed measure- ments were taken at four locations along the curve. The snow- plowable PRPMs appear to have resulted in a smoother speed profile (i.e., there was less variation in speed through the curve) for both directions. In addition, both centerline and edgeline encroachments were reduced significantly. Agent and Creasey (33) investigated the ability of various traffic control measures to delineate horizontal curves so drivers would perceive the curve, slow to an appropriate speed, and then receive guidance through the curve. A before-and- after analysis was carried out to test the performance of PRPMs, transverse pavement stripes, rumble strips, post delin- eators, and chevron signs. Speed and encroachment (centerline and edgeline) data were taken at all sites before and after installation, and a before-and-after crash analysis was per- formed at some of the sites. The after data were taken more than 1 year after installation of the traffic control measures. PRPMs were applied at two regular curves and two S-configurations, each consisting of two 90-degree curves. The PRPMs were applied in pairs along the centerline at inter- vals 10 or 20 ft (3 or 6 m) apart, depending on the location. Two sites—one S-curve and one normal curve—had only PRPMs and no other countermeasures installed. The other two sites had PRPMs and other traffic control measures as well. At the two S-curves with only PRPMs, average speeds were lowered from 23 to 20 mph (37 to 30 km/h) and 25 to 23 mph (40 to 37 km/h), depending on the approach, during the day. 19 At night, average speeds were lowered from 23 to 20 mph (37 to 30 km/h) and from 24 to 22 mph (38 to 35 km/h) after installation of the PRPMs, for the different approaches. These reductions were statistically significant (p < 0.01). The per- centages of encroachments were reduced from 44 to 22 per- cent and from 13 to 7 percent during the day and from 52 to 18 percent and from 8 to 7 percent at nighttime. There was no mention of tests of significance for encroachment data. At the regular curve with only PRPMs installed, average night speeds were reduced from 30 to 27 mph (48 to 43 km/h) on one approach and from 30 to 24 mph (48 to 38 km/h) on the other approach after installation of the PRPMs. These reduc- tions were significant (p < 0.01). Encroachments increased on one approach from 22 to 26 percent and decreased on the other approach from 32 to 29 percent. There was no mention of the radii of the curves or of the tests of significance for encroachment data. The effects of a number of commonly used curve delin- eation treatments on vehicle speed and placement were examined in a study by Zador et al. (21). Treatments were implemented at 51 rural two-lane highway sites. Sites with chevrons, post-mounted delineators, and raised pavement markers were compared with unmodified control sites. Obser- vations were taken at each modified and control site several weeks before and several weeks after the modifications were put in place. Speeds and vehicle placement were taken 100 ft (30 m) before and 100 ft (30 m) into the curve. Of the 51 sites, 12 used standard 44 Stimsonite PRPMs installed on both sides of the double yellow centerline. The markers were usually spaced 80 ft (24 m) apart. Along sharper curves, where three markers could not be seen at one time, the markers were spaced 40 ft (12 m) apart. The results indi- cated that the PRPMs caused the largest shift from the cen- terline as compared with the other countermeasures and the control condition. In advance of the curve 100 ft (30 m), the mean displacement was approximately 0.4 ft (12 cm); 100 ft (30 m) into the curve, the mean displacement was approxi- mately 0.7 ft (21 cm). In comparison, when chevrons were used, the displace- ment from the centerline was less. With post-mounted delin- eators, vehicles moved toward the centerline. Nighttime mean vehicle speeds were increased by approximately 0.68 mph (1.1 km/h) with the use of raised pavement markers; daytime mean vehicle speeds increased by a similar amount. In a nighttime study comparing the impact of PRPMs sup- plementing the existing centerline with post-mounted delin- eators, Krammes and Tyer (34) determined that drivers placed their vehicles further from new PRPMs than from older post- mounted delineators and made fewer encroachments on the adjacent lane. Two factors are operating here: (1) the new PRPMs would be more conspicuous and (2) the new PRPMs would be placed closer to the traveled lane than the post- mounted delineators would. With respect to changes over time, speeds were significantly higher with the new PRPMs than with PRPMs that had been

in place for 11 months, suggesting that drivers had longer pre- view distances and therefore were comfortable with higher speeds. In a recent daytime study, Hammond and Wegmann (35) examined the effects on PRPMs at two spacings on curves by measuring changes in speed and encroachment distances into the opposing travel lane. Two minor arterial sites were chosen for this study. Six data points were collected for each vehicle, including speeds at the beginning, middle, and end of the curve, as well as the distance of opposing-lane encroachment. Levels of encroachment were categorized on a scale of zero to eight, with zero being low and eight being high. A shift in one level of encroachment translates into a 4-in. (10-cm) change of the vehicle travel path. Speeds and encroachment distances were measured before and after the installation of the PRPMs at a 40-ft (12-m) spac- ing and again after the installation of additional PRPMs, cre- ating a 20-ft (6-m) spacing, for a total of 3 experimental con- ditions. Stimsonite LifeLite 88A PRPMs were placed in pairs immediately on either side of the painted centerline. Speed variance was not found to be affected by the pres- ence or spacing of the PRPMs. However, effects on encroach- ment were statistically significant. At Site 1, the control con- dition, the 40-ft (12-m) spacing, and the 20-ft (6-m) spacing yielded levels of encroachment of 4.0, 3.0, and 2.8, respec- tively. At Site 2, the control condition, the 40-ft (12-m) spac- ing, and the 20-ft (6-m) spacing resulted in levels of encroach- ment of 2.7, 1.9, and 1.3, respectively. 2.3.3.2 Driver Speed, Visual Workload, and Lane Position Relative to PRPM Spacing on Curves As discussed above, Hammond and Wegmann (35) com- pared operating speeds and encroachment distances for PRPMs spaced 40 ft (12 m) and 20 ft (6 m) apart. The authors found no significant difference between the 40-ft (12-m) and 20-ft (6-m) spacings in terms of operating speed or centerline encroachments. Blaauw (36) examined drivers’ observation strategy and per- formance on road sections with various delineation arrange- ments. Delineation treatments included PRPMs on edgelines and centerlines at spacings of 40, 80, and 118 ft (12, 24, and 36 m) on straight and curved (radius 656 ft [200 m] and radius 3,280 ft [1,000 m]) sections. A visual occlusion technique was used to determine changes in visual strategy as a func- tion of road delineation. Drivers were equipped with glasses with lenses that could be changed from translucent to opaque almost instantaneously. These glasses allowed them half a second to look at the road on the press of a control switch. The researchers found that total observation time increases and driving performance deteriorates when less delineation infor- mation is present per unit of road length. This finding was par- ticularly striking on the 656-ft (200-m) radius curve where the 20 40- and 80-ft (12- and 24-m) spacing distances lead to speed reductions and lane errors. The authors recommend a mini- mum spacing of 80 ft (24 m) on tangents and 40 ft (12 m) on curves. These results provided a basis for guidelines for the use of PRPMs recommended to FHWA (5). 2.3.3.3 Driver Speed, Lane Position, and Encroachments at Hazardous Locations In 1982, a study was carried out by 12 state highway agen- cies to evaluate the effectiveness of raised pavement mark- ers at hazardous locations (31). Among the test sites were rural curves on two-lane roadways, narrow bridges, stop approaches, through approaches, two-lane sites with left-turn lanes, interchange gores, four- and six-lane undivided sites, multilane divided highway sites, and four- to two-lane sec- tions. Several locations were tested for each site type. The report confirmed that PRPMs provide improved night- time pavement delineation when compared with and used in conjunction with conventional paint stripes. For rural two- lane curves, the report recommended that the double yellow centerline be delineated with one row of PRPMs between the two centerlines, with PRPM spacings of 80 ft (24 m) on curves up to 3 degrees, 40 ft (12 m) on curves between 3 and 15 degrees, and 20 ft (6 m) on curves with more than 15 degrees of curvature. Visual observations indicate that two markers may be needed to provide adequate delineation for locations with curves in excess of 20 degrees. The mixture of centerline and edgeline markings appeared confusing at some sharp curves. The study determined that PRPMs can significantly reduce instances of erratic maneuvers of vehicles through painted gores at exits and bifurcations. This finding was true whether or not overhead lighting was present. The study recommended that PRPMs be introduced slightly in advance of the highway problem area to prepare motorists for the guidance technique that is to be encountered. 2.3.3.4 Driver Speed and Lane Position in Relation to PRPM Spacing on Tangents and Ramps Optimal spacings for PRPMs along tangent sections and on interchange ramps of Interstate highways in Ohio were determined by Zwahlen (29) (see Table 2-8). The first step in this process was to predict the illumination reflected back to the driver’s eyes on the basis of the following: • Headlight output; • Geometry with respect to the PRPM, headlight, and driver eye positions; • Photometric qualities of the PRPM; and • Transmissivity of the atmosphere.

It was assumed that the headlight output was less than 100-percent efficient (how much less was not stated). Also, to account for wear during the life cycle, the specific inten- sity value of the PRPM was assumed to be 50 percent of its new value. Once an illumination threshold for human observers of 98-percent probability of detection was reached, the devices were assumed to be visible. A rain intensity of 1 in. per hour was assumed, since the probability of having a greater rainfall than this within a 30-day period decreases rapidly (e.g., a rainfall of 2.6 in. per hour can be expected once every 25 years). The theoretical calcu- lations indicated that the Stimsonite PRPMs would be visi- ble in rain at 1 in. per hour at a distance of 480 ft (147 m). A model of driver lane position standard deviation on tan- gent sections was then used to predict lane position deviation in relation to the number of PRPMs visible. Once there were four or more delineation devices visible, the model predicted little change in standard deviation. Given a visibility of 480 ft (147 m), to have four devices visible requires a spacing of 120 ft (36 m). On ramp sections in Ohio, the radius of curvature is typi- cally 240 ft (73 m) corresponding to a 24-degree curve. An assumption was made that a solid body of grass or snow 1 to 2 ft (0.3 to 0.6 m) high existed on the inner edge of the ramp curve, limiting the driver’s view ahead. Given this geome- try, the illumination distance for PRPMs will be best if placed on the outer edgeline of the pavement. Low-beam headlights provide the most light to the right; therefore, the left edge of a left curve has the shortest illumination distance—115 ft (35 m). This is about one-fourth of the visibility available on a tangent section. To have four delineators in view in the 115-ft (35-m) distance, an optimal spacing of 25 ft (7.6 m) was selected. The theoretical analysis led to field testing with 11 young subjects in wet and dry conditions. To better replicate the illu- mination levels for PRPMs that have been in service for some time, the new reflectors were cut in half. Vehicle speed and lane position were measured. Conditions tested included spac- ings both wider and narrower than the predicted optimum: 21 • Tangent sections included no PRPMs and PRPMs at spacings of 60, 120, or 240 ft (18, 37, or 73 m). • Ramp sections included no PRPMs and PRPMs at spac- ings of 12.5, 25, or 50 ft (4, 8, or 15 m). On tangent sections, there was a slight but consistent shift of about 5 in. (13 cm) toward the right edgeline for PRPM spacing of 60 ft (18 m) compared with 120 ft (36 m). No sta- tistically significant effects on vehicle speed were found. The study concluded that a 120-ft (36-m) spacing should be rec- ommended on tangent sections. The slight improvement in lane positioning for the 60-ft (18-m) spacing was not felt to justify the doubling in cost of installation. Statistical analysis showed no significant difference in speed or lane positioning related to the presence or spacing of PRPMs on ramp sections; thus, the placement of PRPMs on the outer edgelines of cloverleaf interchange ramps was not recommended. Zwahlen’s study (29) suggests that the problem with PRPMs on very sharp curves (i.e., with a radius curve 240 ft [73 m] or less) is the lack of preview distance (i.e., lacking a preview distance of 120 ft [37 m] or shorter). Related work on chevron spacings (37) used a paradigm in which subjects viewed curves with up to 12 equally spaced chevrons in laboratory conditions that simulated light levels and a size equivalent to 90-degree curves with typical radii seen at night. The subject task was to determine whether the curve viewed was sharper or gentler than a standard curve. Results showed that perfor- mance reached a plateau when four or more equally spaced chevrons were used. Chevrons can be seen considerably fur- ther than PRPMs because of their orientation, and there is no reduction in visibility with rain. Consequently, chevrons seem to be preferable to PRPMs on sharp curves. 2.3.3.5 Driver Response to PRPMs in Deceleration Lanes and at Gore Areas One study examined the impact of PRPMs at gore areas (see Table 2-9). Mullowney (30) found that six out of nine TABLE 2-8 Summary of literature on driver performance and PRPMs (tangent sections) Study Ref. /Location Installation Location PRPM Spacing Effect on Speed Effect on Lane Position Other Zwahlen (1987) (29) Ohio (night testing only) Tangent lane line 59, 121, 240 ft (18, 37, 73 m) No significant difference 5-in. (13-cm) shift away from centerline None Zwahlen (1987) (29) Ohio (night testing only) Ramp edgeline 13, 27, 50 ft (4, 8, 15 m) No significant difference No significant difference Placement of PRPMs on outside edgeline not recommended (no significant difference)

sites where PRPMs had been implemented had statistically significant reductions in vehicles that cut through the painted gore. When PRPMs were placed on the lane and edgeline of deceleration lanes, two out of three sites showed a significant increase in early entry into the deceleration lane. 2.3.3.6 Driver Response to PRPMs at Narrow Bridges A before-and-after analysis of vehicle speed and lateral placement at 18 narrow bridge approach sites was conducted by Bowman and Brinkman (32) (see Table 2-10). The coun- 22 termeasures evaluated were combinations of advance warn- ing signs, pavement markings, PRPMs, roadside delineators, object markers, and adhesive delineators. Measurements of vehicle speed and lateral placement were made using FHWA’s fully automated Traffic Evaluation System (TES). These countermeasures did not result in statistically significant changes in the mean speeds at the p < 0.1 level. However, the countermeasures significantly reduced speed variation when all the vehicle types and time periods were analyzed together. In the Niessner (31) report discussed above, for narrow bridges on rural two-lane roads, a PRPM spacing of 80 ft (24 m) decreasing to 40 ft (12 m) approaching the bridge TABLE 2-9 Summary of literature on driver performance and PRPMs (freeway exits) Study Ref. /Location Site Type Installation Location PRPM Spacing Effect on Speed Effect on Lane Position Other Deceleration lanes Gore, lane line, edgeline 40 ft (12 m) lane line, 40 ft (12 m) edgeline Two of the three sites exhibited a significant increase (p < 0.03, p < 0.01) in early entry into deceleration lane Not determined None Mullowney (1982) (30) New Jersey (night only) Painted gore Gore 20 ft (6 m) Six of nine sites experienced statistically significant reductions (p < 0.02 or less) in cars that cut through painted gore Not determined None Zwahlen (1987) (29) Ohio (night only) Interchange ramps 240 ft (73 m) radius Edgeline 13, 26, 50 ft (4, 8, 15 m) No significant difference No significant difference Placement of PRPMs on outside edgeline not recommended (no significant difference) Study Ref. /Location Installation Location PRPM Spacing Effect on Speed Effect on Lane Position Niessner (1984) (31) USA Edgeline, centerline 80 ft (24 m) decreasing to 40 ft (12 m) Two sites: 1. Significant reduction in nighttime 85th percentile speeds—no significant difference in daytime speeds 2. Vehicle speed at night increased (no mention of daytime speeds) Two sites: 1. Moderate and severe encroachments over the centerline were reduced for both daytime and nighttime 2. No significant difference for daytime or nighttime Bowman and Brinkman (1988) (32) USA Centerline (with other counter- measures) Speeds reduced Unknown (p < 0.1) No significant difference TABLE 2-10 Summary of literature on driver performance and PRPMs (narrow bridges)

resulted in a significant reduction in the nighttime 85th per- centile speeds. There was no significant difference in day- time speeds. Encroachments over the centerline were also reduced significantly for both the daytime and nighttime. The report recommended that PRPMs be placed on both the edgeline and centerline to delineate the decrease in pave- ment width. 2.3.4 Summary Five studies found that PRPMs were associated with fewer encroachments into the adjacent lane on horizontal curves (30, 31, 33, 34, 35). One of these studies examined sites with and without lighting and found that while encroachments were significantly reduced at both sites, they were reduced more at the site with lighting (30). This find- ing confirms the findings of other studies showing that encroachments were reduced during the day as well as at night after the placement of PRPMs on curves. Two studies found that drivers moved away from the PRPMs (27, 34). One study found that, at one site, lane position variability decreased significantly (31). 23 With respect to speed, findings were very mixed, with two studies finding smoother speed profiles through curves (29, 30 night only at one sight), one study finding one site with no significant difference in speed after PRPM application (35), one study finding a significant reduction in 85th percentile speed at one site (31), three studies finding significant increases in speed (27, 31 at one approach only, 34), and one study finding a reduction in speeds both day and night (33). One study that examined the effects 11 months after instal- lation of PRPMs found that speeds were lower, possibly because the lower reflectivity reduces preview distance (34). At night, better delineation may induce higher speeds, par- ticularly on tangents and large radius curves, and possibly on smaller radius curves where only centerline (not edgeline) PRPMs are implemented. This possibility has not been ade- quately studied, but is likely given the results of studies showing that improved delineation (i.e., higher contrast lane striping) was associated with higher speeds. One study found that applying PRPMs in deceleration lanes resulted in drivers entering the deceleration lane earlier (30). This study also found that using PRPMs in gore areas reduced the frequency of encroachments to the gore areas.

Next: Chapter 3 - Data Collection and Preparation »
Safety Evaluation of Permanent Raised Pavement Markers 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 518: Safety Evaluation of Permanent Raised Pavement Markers examines the safety performance of snowplowable permanent raised pavement markers on two-lane roadways and four-lane freeways.

  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!