Effect of Built Environment Factors on Pedestrian Safety in Portuguese Urban Areas

Round 1

Reviewer 1 Report

The aim of this paper is to identify the main factors that influence the pedestrians’ safety in an urban context by taking into account variables related to built and road environment in addition to those concerning traffic flows. The paper is clear and well written, and I believe it is mostly ready for publication. The methodology is fully described and the discussion of the obtained results are convincing. My recommendation is minor revisions and it consists only on a suggestion. In section 2, the authors provide a description of all the groups of factors except that related to “Built environment” factors. For completeness, the authors should briefly describe also this group of factors.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

I appreciate the chance to review this paper about pedestrian safety in Portugal. I appreciate the work that went into this paper, but I have serious concerns about the soundness of the methodology. 

I have two main concerns:

First, the choice of an entire roadway as the dependent variable, rather than a segment of a roadway or an intersection, necessarily means an aggregation of potentially important characteristics into the variables, unless all of these roadways are uniform along their entire length - something of which I'm skeptical. While I do not doubt that the distance between crosswalks could be positively associated with pedestrian crashes, what does this variable actually represent -- the max distance between crosswalks, the average distance between crosswalks, or something else? There are also no count variables for things we know or suspect to impact pedestrian crash risk, including number of driveways and whether or not an intersection is controlled or uncontrolled along the length of the roadway.

Second, while I understand that you are using this model for exploratory purposes, 40 observations is a very low number when you are testing so many variables. 

The GEE model seems more sound from a higher-level, particularly with adjusted exposure, but faces the same issues with the construction of the variables. How can a single exposure variable be reasonably assumed to apply to the entire length of a roadway? You are asking a lot of faith in modeling to be able to accurately represent variation based on a very limited set of variables. 

One way to address these limitations would be to reconstruct your database as a set of segments or intersections and rerunning the model as a zero-inflated negative binomial model instead.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

See the attached file

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

I appreciate that you addressed most of the comments from the first round. I still believe that this type of work is better done in a systemic fashion, like through a safety performance function, which would treat every street segment (or intersection) as its own observation to give more robust findings about the relationship between the independent variables and crashes. That said, I can appreciate that you are trying to establish the theoretical importance of these variables, rather than use them to predict crash counts, and future research can explore these findings more deeply. I recommend adding the idea of future safety performance functions considering these variables in your limitations/conclusions. The following reference would be a good addition:

Thomas, L., Lan, B. Sanders, R. L., Frackelton, A., Gardner, S., & Hintze, M. (2017) Changing the Future? Development and Application of Pedestrian Safety Performance Functions to Prioritize Locations in Seattle, WA. Transportation Research Record, 2659, 212-223. DOI: 10.3141/2659-23

I also encourage you to change the word "accident" to "crash" or "collision".

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

See the attached file.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf