The Impact of Pedestrian Crossing Flags on Driver Yielding Behavior in Las Vegas, NV

Round 1

Reviewer 1 Report

The abstract reads well, however need to remove the statistical results/numbers as well as describe in its end area (s) where the study results could be utilized (e.g. in the area of walking technologies/applications). The study introduction lacks the depth and breadth of a coherent overview of recent studies on walking and driver behavior interventions (at least); and perhaps technologies/apps that are used by walkers and/or drivers and support or inhibit their behavior while on the road. This is crucial for a study of this type for the readers to fully appreciate the proposed study idea. Please check the following references:  

Asimakopoulos, S., Asimakopoulos, G & Spillers, F. (2017). Motivation and user engagement in fitness tracking: heuristics for mobile healthcare wearables. Informatics - Special Issue Smart Health 2016, 4 (1), 5.

Harrison, D.; Marshall, P.; Bianchi-Berthouze, N.; Bird, J. Activity tracking: Barriers, workarounds and customization. In Proceedings of the 2015 ACM International Joint Conference on Pervasive and Ubiquitous Computing, Osaka, Japan, 7–11 September 2015

Gouveia, R.; Karapanos, E.; Hassenzahl, M. How do we engage with activity trackers? A longitudinal study of Habito. In Proceedings of the 2015 ACM International Joint Conference on Pervasive and Ubiquitous Computing, Osaka, Japan, 7–11 September 2015.

Please provide table (s) with the participant demographics as well as the main study results. The results provide a weak/preliminary association between the use of PCFs and driver yielding that can hardly - at this stage - have clear implications for public health/safety as the authors seem to suggest. Perhaps a summary with the main results and implications could clarify this issue. The follow-up discussion is well-written, as well as the emphasis on theoretical and practical study implications. In order to improve the discussion, perhaps a section/paragraph should be described on potential future research and implications for walking - walkers behavior, driver yielding behavior, and technologies to support the proposed study results.

 

The paper is not publishable in its current form.  

 

 

Author Response

Reviewer 1

 

The abstract reads well, however need to remove the statistical results/numbers as well as describe in its end area (s) where the study results could be utilized (e.g. in the area of walking technologies/applications).

 

R: The authors would like to thank the reviewers for their time and feedback for improving the manuscript. We have removed the statistical results from the abstract, leaving only the means. We also added a sentence on more advanced pedestrian safety alternatives. Additionally, further discussion on this is added to the discussion section. The modified abstract sentences are below.

 

Pedestrian crossing flags are a low-tech, low-cost intervention that may improve pedestrian safety at marked mid-block crosswalks. Future research should examine driver fade-out effects and more advanced pedestrian safety alternatives.

 

The study introduction lacks the depth and breadth of a coherent overview of recent studies on walking and driver behavior interventions (at least); and perhaps technologies/apps that are used by walkers and/or drivers and support or inhibit their behavior while on the road. This is crucial for a study of this type for the readers to fully appreciate the proposed study idea. Please check the following references:  

Asimakopoulos, S., Asimakopoulos, G & Spillers, F. (2017). Motivation and user engagement in fitness tracking: heuristics for mobile healthcare wearables. Informatics - Special Issue Smart Health 2016, 4 (1), 5.

Harrison, D.; Marshall, P.; Bianchi-Berthouze, N.; Bird, J. Activity tracking: Barriers, workarounds and customization. In Proceedings of the 2015 ACM International Joint Conference on Pervasive and Ubiquitous Computing, Osaka, Japan, 7–11 September 2015

Gouveia, R.; Karapanos, E.; Hassenzahl, M. How do we engage with activity trackers? A longitudinal study of Habito. In Proceedings of the 2015 ACM International Joint Conference on Pervasive and Ubiquitous Computing, Osaka, Japan, 7–11 September 2015.

R: Thank you for the references. The authors have added a discussion about this in the introduction. It  now reads:

 

Distraction is a primary cause of vehicle and pedestrian related crashes [9,10].  In 2012, it is estimated that one of every five distraction-related motor vehicle crashes involved the use of a hand-held device [9].  Most drivers do not recognize the level to which distractions, particularly using a hand-held device, decrease their attentiveness to controlling their vehicle and monitoring their surroundings [10].  A possible solution to decreasing distraction may be recruiting technology that increases focus on walking and/or driving [11-15].  Wearable technology and apps that track physical activity often encourage users to take a certain number of steps or achieve a target heart rate [11-13].  This goal setting may encourage pedestrians to focus on walking at a brisk pace that would be hard to achieve while looking at a device.  Apps have been designed to track distracted driving [14] or completely disable devices in moving vehicles [15].  When used mindfully, technology may decrease distraction-related motor vehicle crashes.  

 

Interventions aimed at increasing pedestrian safety have had somewhat mixed results. For example, one study aimed to decrease cell phone distractions amongst pedestrians by stenciling safety messages on curbs. While distracted walking decreased initially, the change was not sustained at a four month follow-up [16]. A study in Gainesville, FL, USA used a high-visibility enforcement intervention to increase driver yielding (i.e. increased driver citations, advanced yielding markings and signs, media education campaign, etc.), and found a significant increases in yielding initially which sustained at a four-year follow-up [17]. A meta-analysis of behavioral interventions targeted at child pedestrians concluded that interventions improved children’s pedestrian safety [18].

 

Please provide table (s) with the participant demographics as well as the main study results.

 

R: The authors have provided the full ANOVA table for both dependent variables in the results section – Table 1 and Table 2.

 

 

The results provide a weak/preliminary association between the use of PCFs and driver yielding that can hardly - at this stage - have clear implications for public health/safety as the authors seem to suggest. Perhaps a summary with the main results and implications could clarify this issue.

 

R: The authors have added to the discussion and feel that the implications for public health have been made clear. We have also recommended additional interventions that may enhance findings.

 

The follow-up discussion is well-written, as well as the emphasis on theoretical and practical study implications. In order to improve the discussion, perhaps a section/paragraph should be described on potential future research and implications for walking - walkers behavior, driver yielding behavior, and technologies to support the proposed study results.

 

R: Thank you for this suggestion. The authors have added a paragraph discussing a need for future research related to technology. It now reads:

Further research is needed to identify solutions to increasing driver yielding that do not have a fade-out effect, especially on unsignalized mid-block crossings. While interventions that increase pedestrian visibility and decrease distracted walking and driving are still necessary, technological advancements are one burgeoning crash mitigation method. For example, automatic emergency braking (AEB) systems have the ability to detect pedestrians and decelerate the vehicle to avoid or minimize pedestrian injury. Unfortunately, such technology is not yet widespread and is typically available only in high end cars, though there has been research examining a moderately priced [phone] app based detection system that has the ability to alert drivers without automatic speed deceleration (Mehta & Gupta). While widespread use of such technologies would certainly enhance pedestrian safety, it is important to note, that pedestrian detection is much more efficient at lower travel speeds, with most tests being conducted at speeds of around 20mph. Continued efforts to implement and improve AEB systems are warranted. 

Author Response File: Author Response.docx

Reviewer 2 Report

The article is very clear. It has a very good purpose. But it presents a heavy gap described by the authors. Only two pedestrian mid-block crosswalks were examined.

Some changes are required:
1. Insert a table with the data collected on site.
2. Insert a table with the results of the ANOVA test.
3. Insert the graph showing that the data is similar to a normal distribution.
4. Insert a photo of the test sites while the pedestrian crosses and the cars stop. This increases the attractiveness of the paper.
5. Briefly describe if there are other objects besides PCFs.
6. Discuss whether the ANOVA test can be done on a restricted sample.

Author Response

Reviewer 2

Some changes are required:
1. Insert a table with the data collected on site.

R: Thank you for the feedback intended to improve the manuscript.  There were 3 main variables collected on site, which are laid out in the materials and methods. We have attempted to clarify this, but did not feel that a table made sense given the lengthiness of the verbiage. Please indicate if you would still prefer this in a table.

Insert a table with the results of the ANOVA test.

R: We have added the full anova results for both tests in the results section.

Insert the graph showing that the data is similar to a normal distribution.

R: We have added the histograms for both dependent variables.

Insert a photo of the test sites while the pedestrian crosses and the cars stop. This increases the attractiveness of the paper.

 

Briefly describe if there are other objects besides PCFs.

R: The authors have added a discussion about potentially utilizing technology to auto detect pedestrians. The below paragraph was added to the discussion.

Further research is needed to identify solutions to increasing driver yielding that do not have a fade-out effect, especially on unsignalized mid-block crossings. While interventions that increase pedestrian visibility and decrease distracted walking and driving are still necessary, technological advancements are one burgeoning crash mitigation method. For example, automatic emergency braking (AEB) systems have the ability to detect pedestrians and decelerate the vehicle to avoid or minimize pedestrian injury. Unfortunately, such technology is not yet widespread and is typically available only in high end cars, though there has been research examining a moderately priced [phone] app based detection system that has the ability to alert drivers without automatic speed deceleration (Mehta & Gupta). While widespread use of such technologies would certainly enhance pedestrian safety, it is important to note, that pedestrian detection is much more efficient at lower travel speeds, with most tests being conducted at speeds of around 20mph. Continued efforts to implement and improve AEB systems are warranted. 

Discuss whether the ANOVA test can be done on a restricted sample.

R: The ANOVA F statistic is based on the fundamental compliance with the normality of the dependent variable, which should normally be distributed in the sampled populations, however, if the group sizes are large, the F statistic behaves reasonably well even with population distributions that are far from normal (Games, 1978).

 

Games, P. A. (1978). A three-factor model encompassing many possible statistical tests on independent groups. Psychological Bulletin, 85(1), 168-182.

 

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The manuscript has been significanlty improved.