Effects of Behavior-Based Driver Feedback Systems on the Speeding Violations of Commercial Long-Haul Truck Drivers

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The paper represents an interesting contribution in the field of real-time coaching programs and feedback delivery (the overall topic of ADAS) for road safety improvement, aimed at investigating the effectiveness of on-line feedback for over-speeding reduction while driving, in a wide sample of truck drivers.

High- and low-risk drivers were identified through a K-metoids cluster analysis carried out on of the kinematic driving variables recorded in 4 weeks of observation. In the following 30 weeks, a visual feedback was delivered to half participants when a risky event happens, whereas the other half of the participants did not receive any feedback. Results showed that both high- and low-risk drivers reduced risky events but the former more than the latter. Besides, comparing high-risk drivers who received the feedback and high-risky drivers who did not receive the feedback, the first group showed the greater reduction in risky behaviors.

The paper is well written and the aims are clearly described, as well as data analysis technique and results. The conclusions seem to be supported by the results, but some potentially relevant limitations may have influenced the results and, for this reason, they should be commented explicitly in the discussion and conclusion sections.

Major points

1) With regard to the theoretical background, some additional recent studies on the effectiveness of feedback for the reduction of speeding should be included in the introduction and the present results should be compared with the previous ones, in order to highlight what is the gap covered by the study.

2) On the basis of Figures 2 and 3, the two high-risk groups did not seem to be equivalent at the baseline (first 4-weeks observation). A direct statistical comparison should be carried out before drawing the conclusion that the difference in trends really depend on the feedback. If the new analysis will confirm that the two groups are not similar at the baseline, this could be taken into account in the discussion.

3) The use of the K-metoids method (instead of the k-means used in other studies following the same logic for identifying participants with risky driving styles) should be justified. The authors should argument if the method chosen could have influenced the distribution of participants in the two groups. Could the use of the k-mean method have allowed a better balance between high- and low-risk participants?

Minor points

4) Lines 273-281: The result that “the speeding event rate of higher-risk drivers decreased even though no feedback was provided” should be more parsimoniously explained as an effect due to participants’ expectations (a sort of placebo effect), if also participants in the non-feedback group knew they were involved in the research (or, simply, that their driving behavior was recorded). In addition, how was the informed consent collected? I can’t find mention to the ethical approval in the paper. (I’m not able to find out if it is mandatory for the journal).

5) The authors stated that the cluster analysis was carried out only on speeding events. Is this also true for the feedback delivery, or was the feedback presented also when other kinds of safety-related events occurred?

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

An interesting and well-written manuscript concerning an investigation of the drivers' speeding event rates of truck drivers in order to evaluate the effects of a safety program in a trucking fleet.

The results of the program are interesting. Two remarks for consideration:

- The K-medoids algorithm was used to group drivers into two distinct clusters based on their speeding event rates during the first 4 weeks (stage 1). Extended analysis for this methodology is needed.

- Mean speeding event rates for feedback group and no-feedback group for high risks drivers are quite different, in Stage 1. Which is the reason for this ?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

 Safety

Manuscript Number: safety-2744516

Manuscript type: Article

Title: Effects of Behavior-based Driver Feedback Systems on Speeding Violations of Commercial Long-Haul Truck Drivers

The paper examines the impact of behavior-based safety (BBS) programs on speeding.

In the frame of this study, the Authors attempt to address three questions:  (1) Do behavior-based safety programs help reduce speeding violations? 83 (2) What is the impact of these programs on speeding violations for different styles of drivers? (3) How do these programs impact speeding violations over time?

Overall, this is an interesting and important topic, which can be a good match for the Safety Journal. No doubt, this paper contributes to road safety but its contribution to the existing literature is marginal.  Based on the introduction and the conclusion, the contribution of this paper to the existing literature is not clear! I do not recognize the innovation in this paper.

However, some issues require attention and revisions before I can recommend publication.

Data and method:

It’a missing the characteristics of the drivers, as well as the demographic and the socio-economic features. I believe that this part of these features is available. In addition to the driver’s experience, how many hours do they work? Where do they drive, intercity or intra-city? It is also important. These features have the potential to affect the results of the study.

Do participants know about the experiment? I believe yes. Especially the group that did not get feedback. The fact that drivers are aware that there is an experiment and monitor their behavior can affect their driving.

How was the speed event defined? Many levels of speed violation can be defined, some very dangerous and some less, so it is important to define it. There is a huge difference between exceeding the speed limit by 10 km/h or by 20 or more?

I do not think that it is a good idea to split the sample into two groups, risky and not risky, because the sample is so small and the speed violation baseline is not the same. My recommendation is to analyze both samples together in addition to the presented analysis.

 

The authors attempt to examine the impact of these programs on speeding violations over time.

To examine the long-term effect of this program on driving behavior “speeding violation in this case” it is important to follow the drivers’ behavior after the experimental period “ Post-experimental period”.  At this stage, drivers will not be aware that experiment investigators continue to monitor their driving.

Please see:

Elias, Wafa. "The effectiveness of different incentive programs to encourage safe driving." Sustainability 13.6 (2021): 3398.

 

Results

Before Figure 1 there is another unnumbered figure that presents the drivers ordered by speeding event rate for the feedback group. If authors attempted to present both figures with and without feedback together, they have to present clearly.

It is interesting to present the findings in Figures 2 and 3 together within the same figure.

 

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have revised the manuscript taking into account the proposed comments. 

I agree with the revised paper and I propose the paper to be published.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have addressed most of my outstanding points in a coherent and well-argued way. Good work! No more comments.