Evaluation of Pedestrian Hybrid Beacons and Rapid Flashing Beacons

CHAPTER 4. DRIVER-YIELDING RESULTS FOR BEACONS PLACED ABOVE OR BELOW CROSSING SIGN IN AN OPEN-ROAD SETTING

INTRODUCTION

For the open-road study, the test conditions were set to determine driver yielding when the beacons were located above or below the warning sign. Both placements were studied at all sites so that a similar driver population would see both treatments. This chapter describes the methodology and results from the open-road study that investigated the effects of the placement of yellow rapid-flashing beacons above or below the pedestrian crossing sign.

Due to the findings documented in this report, FHWA issued another interpretation: Official Interpretation #4(09)-58 (I)—Placement of RRFBs Units Above Sign.⁽³⁾ This permits the placement of the beacons either above or below the crossing warning sign.

Study Overview

When IA-11 was issued in July 2008 for the RRFB, the only position of the beacons described in the document was below the crossing warning sign and above the supplemental downward diagonal arrow plaque.⁽⁴⁾ As described in chapter 3 of this report, the position of the beacons had an effect on drivers' time to detect the presence and direction of crossing pedestrians as well as discomfort glare during nighttime conditions on a closed course. Prior to developing the proposed provisions for incorporating a rapid-flashing beacon traffic control device into the MUTCD, it is important to determine which beacon position is most beneficial from a driver yielding perspective.⁽¹⁾ This study sought to determine if mounting the beacons above the pedestrian crossing sign was more effective in terms of driver yielding than the traditional position below the sign.

Study Objective

The objective of this study discussed in this chapter was to determine benefits of different positions for the RRFBs used with pedestrian crossing signs in an open-road setting. Because the closed-course study presented in chapter 3 indicated that benefits may exist for placing the beacons above the sign, the open-road study investigated if drivers yielded differently to RRFBs placed above versus below the pedestrian crossing sign.

STUDY DEVELOPMENT

Study Sites

Near the conclusion of the closed-course study described in chapter 3, the researchers talked to agency representatives and made requests during professional society meetings, seeking agencies that would be willing to participate in the open-road research. Four agencies volunteered: Aurora, IL; Douglas County, CO; Marshall, TX; and Phoenix, AZ. As a minimum, the agencies were asked to identify at least two locations that either had existing RRFBs below the pedestrian crossing sign that could be moved to the position above the sign or that would allow the beacons to be installed in one position and then moved to the other position after the initial data collection. Table 49 lists the 13 sites included in the study. The average daily traffic (ADT) values were provided by the agencies in Arizona, Colorado, and Texas. Researchers estimated the ADT for the Illinois sites based on 1-h counts made from the video recordings.

Note: Sites are labeled as XX-YY-##, where XX represents the two-letter State code; YY represents the two-letter city code, and ## represents the site number within the city.
NR = No refuge; TWLTL = Two-way left-turn lane.
^aThe distance (ft) to nearest intersection or major driveway is shown in parentheses (measured from the center of the crossing to the center of the nearest driveway/intersection).
^bThis indicates the number of pedestrian crossings per hour during the daytime data collection period when the beacons were located below the crossing sign.
^cSpeed limit varied by approach.
^dSite also includes the following two advance traffic control assemblies: pedestrian crossing (W11-2) warning sign with AHEAD (W16-9P) plaque, and SPEED LIMIT 25 (R2-1) regulatory sign with WHEN FLASHING (S4-4P) plaque and a 12-inch circular beacon that is activated when the pedestrian pushes the pushbutton at the crossing.

Study Assemblies

Examples of the study assemblies are shown in figure 47 and figure 48. The beacons were mounted on a roadside pole to supplement either a W11-2 (pedestrian) or W11-15 (trail) crossing warning sign with a diagonal downward arrow (W16-7p) plaque and located at or immediately adjacent to a marked crosswalk. The flash pattern used at the study sites was the 2-5 flash pattern. Table 50 provides information on installation order along with the dates of the data collection.

Figure 47. Photo. Example of RRFB placed above the sign.

Figure 48. Photo. Example of RRFB placed below the sign.

Existing = RRFB was installed below the sign at site prior to the study.
Activated = Pedestrian-activated yellow circular 12-inch flashing beacons were activated.

Rotation

To account for the possibility that device installation order could affect the results, the RRFB was installed above the sign first in some locations and second in other locations. For the 13 study sites, the RRFB was installed initially above the sign in 3 of the sites and was previously installed or initially installed below the sign at the remaining sites.

DATA COLLECTION AND REDUCTION

Study Periods

The study was conducted between October 2014 and May 2015. Following installation of the device in its initial position, the research team collected after data. Once the after data were obtained, the research team requested that the device be installed in the second position (i.e., RRFBs above the sign were moved below the sign and vice versa). After receiving confirmation that the devices had been moved, the research team collected after data for the second position.

Data were collected primarily during the daytime when vehicles were free-flowing. Because few studies have collected data at night, the research team wanted to obtain some data for nighttime conditions. The characteristics of the beacons and the site may have different impacts on driver yielding during night conditions as compared to daytime conditions. Therefore, nighttime data were collected for one site within each city.

Staged Pedestrian Protocol

The research team used a staged pedestrian protocol to collect driver yielding data to ensure that oncoming drivers received a consistent presentation of approaching pedestrians. Under this protocol, a member of the research team acted as a pedestrian using the crosswalk to stage the conditions under which driver yielding would be observed. Each staged pedestrian wore similar clothing (gray t-shirt, blue jeans, and gray tennis shoes) and followed specific instructions in crossing the roadway. The staged pedestrian was accompanied by a second researcher, who observed and recorded the yielding data on pre-printed datasheets.

Prior to the staged crossing maneuvers, researchers placed markers (either small contractor flags or cones) at the edge of the traveled way at a distance corresponding to the AASHTO SSD value for the posted speed limit at that site; one marker was placed in each direction approaching the crosswalk.⁽³⁸⁾ SSD is 200 ft for 30 mi/h, 305 ft for 40 mi/h, and 360 ft for 45 mi/h. After the study site had been prepared, the researchers followed the predetermined staged pedestrian protocol, which was defined as follows:

1. The staged pedestrian approached the crosswalk as oncoming vehicles approached the SSD marker activating the RRFB.
   
   
2. The staged pedestrian reached the edge of the crosswalk in time to place one foot in the crosswalk (e.g., off the edge of the curb or curb ramp) within approximately 1 s of the approaching driver(s) reaching the SSD marker.
   
   
3. The staged pedestrian waited to cross until approaching drivers yielded or until all approaching drivers had traveled through the crosswalk.
   
   
4. The observer recorded how many motorists did/did not yield as well as how many were in a position to yield for each crossing maneuver. Drivers were considered to be in position to yield if they were upstream of the SSD marker when the staged pedestrian was positioned at the edge of the crosswalk. Each such vehicle that did not yield was counted as was each yielding vehicle. Of the vehicles in a position to yield, a vehicle was considered to be yielding if the driver slowed or stopped for the purpose of allowing the waiting pedestrian to cross. Any vehicles traveling in a platoon behind yielding vehicles were not counted because those drivers did not have the opportunity to make a decision on whether to yield to the pedestrian; therefore, the maximum number of yielding vehicles possible for each crossing maneuver was equal to the number of travel lanes through which the crosswalk passed.
   
   
5. Yielding was observed separately for each direction of vehicular travel because Arizona, Colorado, Illinois, and Texas law is written such that drivers must yield to pedestrians in or approaching their half of the roadway.
   
   
6. The observer noted any unusual events or noteworthy comments for each crossing.
   
   
7. Once the crosswalk was clear (i.e., the approaching vehicle had either stopped or passed through the crossing), the staged pedestrian crossed the street and waited on the sidewalk or roadside until all vehicles visible during that crossing traveled through the crosswalk. After all such vehicles had left the study site, the staged pedestrian prepared for the next crossing maneuver.

The protocol called for the completion of a minimum of 20 staged crossing maneuvers in each direction of travel for a total of 40 crossings. Observation periods were chosen such that vehicle traffic was heavy enough to create frequent yielding situations but not heavy enough for congestion to affect speeds. Data were always collected during daylight and in good weather, avoiding rain, wet pavement, dusk or dawn, or other conditions that could affect a driver's ability to see and react to a waiting staged pedestrian.

A minimum of 40 (and a desired 60) staged pedestrian crossings were collected at each site within each time period during daytime. Because of the length of time needed to collect the crossing, a minimum of 40 staged pedestrians were collected at night.

Driver Yielding

After completing the data collection, researchers entered the crossing data and the site characteristics data from the field worksheets into an electronic database. The average yielding rate for a site was calculated, as shown in figure 49; however, data for individual crossings were used in the statistical evaluation.

Figure 49. Equation. Driver yielding rate.

Table 51 lists the driver yielding rates for each site and beacon position along with the number of staged pedestrian crossings for the nighttime data collection periods. Driver yielding to staged pedestrians at night averaged 68 percent for the above position and 65 percent for the below position.

Table 52 shows similar results for the daytime data collection periods. During the daytime, driver yielding to staged pedestrians averaged 64 percent for the above position and 61 percent for the below position. For several sites, neither beacon position showed a large increase in driver yielding as compared to the other. The range of driver yielding to staged pedestrians at these sites ranged from 19 to 98 percent.

RESULTS

When a driver approaches a pedestrian crossing, the driver either yields and stops (or slows) the vehicle or does not yield to the waiting pedestrian. This binary behavior (yield or no yield) can be modeled using logistic regression. A significant advantage of using logistic regression is it permits consideration of individual crossing data rather than reducing all the data at a site to only one value. For the dataset available within this study, that means that over 1,900 data points could be available (i.e., all the unique staged crossings recorded) rather than only 34 data points (i.e., the number of study sites by number of assemblies and by day or night). For the analyses that focused on comparing the below position to the above position, that means 1,507 data points rather than 26 data points were available. These larger sample sizes could result in finding significant relationships that would not be apparent with a smaller dataset. Additionally, it is possible to utilize random effects to account for site-specific differences since such differences induce a correlation structure in the dataset.

Using logistic regression to model the relationships assumes that the logit transformation of the outcome variable (i.e., yielding rate) has a linear relationship with the predictor variables, which results in challenges in interpreting the regression coefficients. The interpretation of such coefficients is not on the yield rate changes directly but a change in the odds of motorists yielding (odds are defined as the ratio of the number of yielding motorists to the number of non-yielding motorists). The regression coefficients can be transformed and interpreted as odds ratios of different levels of the corresponding independent variable. In other words, a unit change of the independent variable corresponds to a change in the odds of motorists yielding, which is an alternative way to express a change in yielding rate. More details on these types of models can be found in the literature.⁽⁴⁷⁾ All the statistical analyses were performed using R, an open-source statistical language and environment, and two open-source packages for fitting GLMMs.^(48,44,45)

COMPARISON OF BELOW TO ABOVE

Because a previous study that included RRFBs found posted speed limit, crossing distance, and city influenced driver yielding, the initial analyses were also conducted with those variables.⁽³⁵⁾ In addition, a variable to reflect the intersection configuration was included, as preliminary reviews indicated that the number of approach legs may be related to yielding results.

Preliminary modeling revealed a correlation between road type (e.g., number of lanes and median treatment) and speed limit present in the dataset; therefore, only posted speed limit was included in the final model. Models were examined that included other variables, such as total crossing distance; however, the best results were found when the variables shown in table 53 were included. The reference level for a driver yielding in the model was estimated for the following conditions: an above sign during the daytime in Arizona with a three-leg intersection.

From the preliminary review of the results in table 52, it appears that there were only minor, if any, differences between the above and below position for the RRFBs. The results from the GLMM are shown in table 53, and these results support that observation. The results indicate that there were no significant differences between the two beacon locations (p-value = 0.1611).

The day/night variable was significant (p-value = 0.0005), which indicates that there were day/ night differences for this dataset regarding driver yielding. It appears that Illinois had notably lower driver yielding as compared to the base State, Arizona. An adjusted p-value for multiple comparisons is required to make a formal assessment. Texas and Colorado were not different from Arizona. The model also indicates that the driver yielding at the midblock offset configuration was statistically different from the driver yielding at the three-legged intersections. A caution with this finding is offered since there was only one site with a midblock offset configuration in the dataset.

Estimate = Natural logarithm of the ratio = Odds (coefficient level)/Odds (reference level). In the case of reference level, estimate is the log-odds of the average yielding rate at the reference level.
t-value = Conservative estimate of the z-value, which is the standard normal score for the estimate, given the hypothesis that the actual odds ratio equals 1.
p-value: Probability that the observed log-odds ratio is at least as extreme as the estimate, given the hypothesis that the actual odds ratio equals 1.
^aReference level driver yielding in the model is estimated for the following conditions: above, day, Arizona, and three-legged intersection.
^bThese p-values require an adjustment for multiple comparisons if inferences about different yielding rates among States or among configuration are intended.
^cSignificance values are as follows: blank cell = p > 0.10; ~ = p < 0.10; * = p < 0.05; ** = p < 0.01; and *** = p < 0.001.