Review Reports - Effects of Varying Text Message Length and Driving Speed on the Disruptive Effects of Texting on Driving Simulator Performance: Differential Effects on Eye Glance Measures

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper describes an experiment regarding texting while ‘driving’ in a simulator.

Lines 90-97

Caird et al. (2014) report : “Drivers who read and typed texts also tended to decrease their speed and increase the distance to vehicles in front of them, which partially compensated for impairments produced by looking away from the roadway”.

Did you find this result in your experiment?

Lines 122, 123

Why did you choose 60 and 80 mph? Is 80 mph realistic for texting?

Author Response

Reply to reviewer #1:

This paper describes an experiment regarding texting while ‘driving’ in a simulator.

Lines 90-97

Did you find this result in your experiment?

In the present study, the subjects were instructed to drive at either 60 or 80 mph. Thus, we did not encounter any speed changes while texting. The Caird et al findings of reduced driving speed while texting has been included in the discussion of the revised manuscript.

Lines 122, 123

Why did you choose 60 and 80 mph? Is 80 mph realistic for texting?

The driving speeds of 60 and 80 mph were selected with the hope that they reflect the 25th and 75th percentile values for speeds that might commonly be encountered in freeway driving. This has been noted in the revised manuscript.

Reviewer 2 Report

Comments and Suggestions for Authors

Pictures of the simulator set up should be provided along with pictures from the driving scenario itself.

How long was the driving round (time and length)? How many rounds for each driver? Is there any simulator sickness noticed?

Line 138, What is weaving? Is it the weaving maneuver as defined in the Highway Capacity Manual?

The number of participants is relatively low.

The key points section did not mention the impact of increasing the speed and texting.

The contribution of this paper should be explained in light of this research "Simulation of texting impact on young drivers’ behavior and safety on motorways"

The formatting of the paper is so problematic. Figures formatting sometimes is not consistent even within the same figure as in Figure 1. Reference formatting needs to be adjusted.

Comments on the Quality of English Language

There is unneeded spaces between words in many parts of the manuscript.

Author Response

Comments and Suggestions for Authors

Pictures of the simulator set up should be provided along with pictures from the driving scenario itself.

We thank the reviewer for this suggestion. Given the length of the manuscript, however, and the fact that the simulator and the roadway used (straight road, no oncoming traffic) is not particularly unique, we opted to NOT include a photo of the simulator and/or roadway.

How long was the driving round (time and length)? How many rounds for each driver? Is there any simulator sickness noticed?

After several ‘drives’ to accommodate the subject to the driving simulator, each subject drove the simulator on two occasions, one in which the text message length varied and a second occasion in which the driving speed was varied. Each ‘drive’ was approximately 10 minutes in duration. Also, no subject reported experiencing ‘simulator sickness’, largely because this roadway is largely straight. This information has been included in the revised manuscript.

Line 138, What is weaving? Is it the weaving maneuver as defined in the Highway Capacity Manual?

Our use of the term weaving was imprecise and inaccurate. We have corrected the revised manuscript and the term weaving has been replaced by deviations within the lane. We thank the reviewer for this suggestion.

The number of participants is relatively low.

This point also was raised by Reviewer #4. The sample size, N=9, was sufficient to demonstrate statistically significant outcomes and differences between the effects of changing speed versus changing the text message length. Driving performance, e.g., SDLP and eye glance measures, can vary considerably between subjects, but these values actually very stable across test drives for any given subject. Our use of a within-subjects design allowed us to capitalize on this test-retest reliability. We are confident that our sample size did not result in any Type II statistical errors, i.e., failing to identify a treatment effect.

The key points section did not mention the impact of increasing the speed and texting.

The key points section in the manuscript has been revised to reflect that fact that both increasing text message length and increasing driving speed are associated with greater disruption of driving while texting, as assessed by both (1) evaluation of driving videos and (2) SDLP measures.

The contribution of this paper should be explained in light of this research "Simulation of texting impact on young drivers’ behavior and safety on motorways"

The manuscript has been revised to reflect the contribution of our findings with respect to creating more nuanced ‘Don’t text while driving’ messaging. Specifically, such messages would need to be carefully prepared, since one would NOT want to create messaging that suggests that texting while driving is safe as long as the messages are short or the driving speed is slower.

The formatting of the paper is so problematic. Figures formatting sometimes is not consistent even within the same figure as in Figure 1. Reference formatting needs to be adjusted.

The figures have been reformatted to be internally consistent. In addition, the previously unwieldy Figure 2 (six panels) has been split into two figures for better flow and comprehension. In addition the references have been formatted appropriately.

Comments on the Quality of English Language

There is unneeded spaces between words in many parts of the manuscript.

We have tidied up the spacing in the revised manuscript.

Reviewer 3 Report

Comments and Suggestions for Authors

The paper is well-written, and its structure is clearly presented. Tests are adequately described, and the consequences are clear.

My only concern is about the figures. Figures 1 and 3 consist of four sub-figures, and Figure 2 has six separate parts, all with a lot of text within the figure, most of which is repeated in the figure captions and sometimes in the main text. Please consider splitting figures in separate ones (two in one would also do) and avoiding duplications of texts within the figure, in the caption and in the main text body.

Author Response

The paper is well-written, and its structure is clearly presented. Tests are adequately described, and the consequences are clear.

As mentioned above, the previously unwieldy Figure 2 (six panels) has been split into two figures for better flow and comprehension. In addition the references have been formatted appropriately. In addition, we have reduced some of the text in the figures.

Reviewer 4 Report

Comments and Suggestions for Authors

The study addresses an important public health issue by exploring the effects of texting while driving, focusing on both text message length and driving speed. The findings are relevant and contribute to the understanding of distracted driving behavior.

Detailed Comments:

MATERIALS AND METHODS SECTION:

The study has a small sample size (n=9), which may limit the generalizability of the findings. Expanding the sample size would provide more robust results and enhance the statistical power of the analysis.

Did any of the 9 authors of the study participate as an experimental subject?

Participant Demographics: Specify driving experience (years with license, miles or km driving per year,..). And visual conditions: Did the subjects use optical correction?, Was their vision good? Were visual problems ruled out?

The study only includes young adults (18-25 years old), which may not represent the broader driving population. Including participants from a wider age range would help in understanding how texting while driving affects different age groups.

Lines 101-102: you write in the text: "All subjects used smart phones and reported texting more than 80 times per day". It's important to note that the participants in the study have experience with writing text messages (80 times per day). Should the participants' varying levels of experience in writing messages be considered as a potential influencing factor on the results? This should be addressed in the discussion.

Lines 120-121: Are there specific criteria for selecting these words or phrases? How many times were they repeated in each experiment?, specify in the text.

Lines 149-150: How many subjects (trainer scores) evaluated each video?, specify in the text.

Statistical Analyses: Justify why parametric statistics are sometimes used with such a small sample of subjects.

RESULTS SECTION:

In the results section, it includes content that belongs in the discussion section. Consider adjusting the content to correspond to each section appropriately.

DISCUSSION SECTION:

Authors should discuss the results and how they can be interpreted in perspective of previous studies and of the working hypotheses. The findings and their implications should be discussed in the broadest context possible. Future research directions may also be mentioned.

The discussion could be expanded to explore the broader implications of the results, particularly how they might inform public health campaigns and policy decisions related to distracted driving.

Control Conditions: The study lacks a control condition that includes other forms of distracted driving (e.g., phone calls, interacting with in-car systems). Comparing texting while driving with other distractions could provide a more complete understanding of its relative risks.

Include in the discussion as a limitation of the study the driving scenarios.

Future research directions may also be mentioned.

REFERENCES:

Please update the citation format to comply with the journal's ACS style guide for all references. Refer to: https://mdpi-res.com/data/mdpi_references_guide_v9.pdf

Lines 398-399: Please, revise Reference n. 2 details (Araujo, 2017). I have not been able to find it and it does not appear in Pubmed or Google Scholar.

Line 405: Verify the website link.

Line 422: Verify the website link.

FIGURES:

Comments:

Figure 1 Upper Panel (left and right): Use the same scale for both graphs, on the y-axis scale (mean driving score) to facilitate comparisons and identify trends.

Figure 2: The caption for the figure is not placed in the correct location.

Figure 4 Use the same scale for both graphs, on the x and y-axis scale to facilitate comparisons and identify trends.

Given the small number of subjects in the study, using box plots would provide more accurate graphs, as they display the median and the interquartile range.

Author Response

Detailed Comments:

MATERIALS AND METHODS SECTION:

Please see reply to Reviewer #2.

Did any of the 9 authors of the study participate as an experimental subject?

Several of the authors did in fact participate as subjects in the present study. Their participation was IRB-approved and was completed prior to their involvement (nearly one year later) in analyzing the data and writing the manuscript.

Participant Demographics: Specify driving experience (years with license, miles or km driving per year,..). And visual conditions: Did the subjects use optical correction?, Was their vision good? Were visual problems ruled out?

All subjects had valid driving licenses, had been licensed for 3-9 years (average: 5.6), and drove 1500 - 4500 miles/week (average: 3640). Subjects with glasses or contact lenses also wore those devices during the driving simulator tests. This information is included in the revised manuscript.

It is correct that only young drivers were studied. Clearly, texting while driving is a problem that is NOT restricted to young drivers; indeed, when compared to younger drivers, older drivers have been shown to be more adversely affected by texting while driving (Rumschlag et al, 2014). Future studies examining the variables of text message length and driving speed in older drivers are planned.  This point is noted in the revised manuscript.

Lines 101-102: you write in the text: "All subjects used smart phones and reported texting more than 80 times per day". It's important to note that the participants in the study have experience with writing text messages (80 times per day). Should the participants' varying levels of experience in writing messages be considered as a potential influencing factor on the results? This should be addressed in the discussion.

The Reviewer is quite correct that texting experience can be a significant factor. Regarding our assessment of texting experience of the subjects, the subjects were asked to self-identify their texting experience. At that time, the category of ‘greater than 80 times per day’ was the highest incidence category available to choose, and was selected by all of the subjects. Our primary concern was to be sure that all of our subjects were well-experienced texters. In future studies, we will expand the experience assessment question.

Lines 120-121: Are there specific criteria for selecting these words or phrases? How many times were they repeated in each experiment?, specify in the text.

The phrases used for texting were created simply because they reflected different levels of key strokes. All subjects were exposed to all four text messages over the course of the testing sessions.

Lines 149-150: How many subjects (trainer scores) evaluated each video?, specify in the text.

Analyses of the videos for both driving performance and eye glance measurements were conducted by 3-5 trained scorers who were blinded regarding the treatment condition. Inter-rater reliability was strong, with correlation coefficients of 0.85-0.96 between individual raters. This information is included in the revised manuscript.

Statistical Analyses: Justify why parametric statistics are sometimes used with such a small sample of subjects.

The data collected (SDLP, eye glance number and duration) were interval in nature and were normally distributed. Thus parametric statistics were deemed appropriate for the analyses.

RESULTS SECTION:

In the results section, it includes content that belongs in the discussion section. Consider adjusting the content to correspond to each section appropriately.

It is not clear specifically what parts of the text the Reviewer would like to have moved from the Results into the Discussion section.

DISCUSSION SECTION:

The discussion could be expanded to explore the broader implications of the results, particularly how they might inform public health campaigns and policy decisions related to distracted driving.

The Discussion has been expanded to further address one of the most novel aspects of the present study, in particular, the observation that the longest eye glance away from the road is most likely to occur late in the ‘texting bout’, and the challenges that come with sharing that message with the public without.

Include in the discussion as a limitation of the study the driving scenarios.

Future research directions may also be mentioned.

The Reviewer is correct that we did not study any other driving distractions as a point of reference. This is true. We are currently studying the effects of cognitive distractions on driving performance.

REFERENCES:

Please update the citation format to comply with the journal's ACS style guide for all references. Refer to: https://mdpi-res.com/data/mdpi_references_guide_v9.pdf

Lines 398-399: Please, revise Reference n. 2 details (Araujo, 2017). I have not been able to find it and it does not appear in Pubmed or Google Scholar.

Line 405: Verify the website link.

Line 422: Verify the website link.

The references have been tidied up per the comments of this Reviewer.

FIGURES:

Comments:

Figure 1 Upper Panel (left and right): Use the same scale for both graphs, on the y-axis scale (mean driving score) to facilitate comparisons and identify trends.

Figure 2: The caption for the figure is not placed in the correct location.

Figure 4 Use the same scale for both graphs, on the x and y-axis scale to facilitate comparisons and identify trends.

Given the small number of subjects in the study, using box plots would provide more accurate graphs, as they display the median and the interquartile range.

As mentioned in the comments to Reviewer #3,, the previously unwieldy Figure 2 (six panels) has been split into two figures for better flow and comprehension. In addition the references have been formatted appropriately. In addition, we have reduced some of the text in the figures. The plots include the parametric values of Mean + SEM, since the data were interval in nature and were relatively normally distributed.

Reviewer 5 Report

Comments and Suggestions for Authors

The research has some implications in driving safety, but the following major problems remain.

1. The sample size of the study was too small. As a study in traffic safety, the sample size has a significant impact on the results. Only nine volunteers were selected for the manuscript, which makes it difficult to draw convincing conclusions.

2. Insufficient originality. There is a lot of research on driving behavior safety, and the articles fail to reflect improvements in variables, algorithms or models.

3.Quality of Presentation needs improvement. For example, the positions of Figures 3 and 4 in the manuscript are not uniform, and the font and line spacing of the second paragraph in the Discussion section are rather confusing.

Comments on the Quality of English Language

Some of the sentences in the text are too long and could be appropriately streamlined for description.

Author Response

The research has some implications in driving safety, but the following major problems remain.

The sample size of the study was too small. As a study in traffic safety, the sample size has a significant impact on the results. Only nine volunteers were selected for the manuscript, which makes it difficult to draw convincing conclusions.

This point also was raised by Reviewer #2. The sample size, N=9, was sufficient to demonstrate statistically significant outcomes and differences between the effects of changing speed versus changing the text message length. Driving performance, e.g., SDLP and eye glance measures, can vary considerably between subjects, but these values actually very stable across test drives for any given subject. Our use of a within-subjects design allowed us to capitalize on this test-retest reliability. We are confident that our sample size did not result in any Type II statistical errors, i.e., failing to identify a treatment effect.

Insufficient originality. There is a lot of research on driving behavior safety, and the articles fail to reflect improvements in variables, algorithms or models.

As mentioned above, the figures have been reformatted to be internally consistent. In addition, the previously unwieldy Figure 2 (six panels) has been split into two figures for better flow and comprehension. Also, the positions of Figures 3 and 4 in the manuscript are now uniform. Finally, the font and line spacing of the Discussion have been corrected.

Comments on the Quality of English Language

Some of the sentences in the text are too long and could be appropriately streamlined for description.

The text has been reviewed and streamlined where possible.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Thank you for responding to my remarks. The revised version is OK.

Author Response

Thank you for responding to my remarks. The revised version is OK.

Thank you for your previous review.

Reviewer 2 Report

Comments and Suggestions for Authors

I still with the opinion that the inclusion of the driving simulator set up and the scenario design will be of higher benefit for the reader and will help understand the atmosphere when the drives took place. My understanding is that there is no limit on the number of figures or the number of pages.

The study should clearly note the following limitation:

- further research should include more drivers covering various demographics and experience levels. Nine drivers are not enough to generalize the findings.

- future research should consider the impacts of surrounding road infrastructure (i.e., segments vs intersections) as suggested in " Location-based analysis of car-following behavior during braking using naturalistic driving data"

Author Response

The manuscript now includes Figure 1, which is a photo of the driving simulator and the image seen by subjects.

The study should clearly note the following limitation:

- further research should include more drivers covering various demographics and experience levels. Nine drivers are not enough to generalize the findings.

The manuscript now includes a note indicating that future studies examining the variables of text message length and driving speed in older drivers and across various texting experience levels are planned.  

The revised manuscript now includes a statement indicating that future research should consider the impacts of surrounding road infrastructure such as intersections and ambient traffic.

Reviewer 3 Report

Comments and Suggestions for Authors

My former comments have been adequately addressed.

Author Response

My former comments have been adequately addressed.

We thank you for your review.

Reviewer 4 Report

Comments and Suggestions for Authors

Since I don’t have access to the authors’ responses to other reviewers regarding issues I have also raised, I kindly request that my previous comments and requests be addressed as thoroughly as possible.

Author Response

We apologize for this oversight. Below are the replies that were provided to other reviewer regarding these same issues:

Please see reply to Reviewer #2.

From Reviewer #2 - The number of participants is relatively low.

Raised initially by Reviewer #4 - Given the small number of subjects in the study, using box plots would provide more accurate graphs, as they display the median and the interquartile range.

Raised initially by Reviewer #3: The formatting of the paper is so problematic. Figures formatting sometimes is not consistent even within the same figure as in Figure 1. Reference formatting needs to be adjusted.

Reviewer 5 Report

Comments and Suggestions for Authors

No more comments.

Author Response

No more comments.

We thank you for your review.

Round 3

Reviewer 2 Report

Comments and Suggestions for Authors

Reviewer 4 Report

Comments and Suggestions for Authors

After reviewing the study, I find its approach interesting. However, it notably fails in its methodology and objectivity. Among other issues, it should involve more participants, use appropriate statistical methods, and improve the results and discussion.

I have concerns about the reliability of the results obtained from the study, which only involved 9 participants. There is a significant possibility of biases in the study that could invalidate the results. For example, some of the subjects analyzed were also the study’s authors, which could introduce self-selection bias or participation bias, affecting the objectivity and validity of the results. Additionally, there are inconsistencies in the sample characterization, such as the amount of driving experience and miles driven per week. The study only includes young adults (18-25 years old), yet the participants have had licenses for 3-9 years on average and drove 1500-4500 miles per week on average (3640 miles).

They need to adequately address the issues raised in the review, and the manuscript still has significant gaps that make it unsuitable for publication in its current form.