Dear authors:

The reviewers clearly state that this manuscript is a valuable contribution to the literature supporting the promotion of rip current safety. I am generally pleased with how you addressed the reviewers’ comments and believe that the revisions have resulted in an improved manuscript.

I have a few additional minor comments that I would like you to address before uploading the revised manuscript for final review by me.

1) AR3 & AR4: You state that in response to these concerns, you have expanded Section 4.2, which discusses the implications for beach safety. Given the possible implication of the highlighted issue on your results, it seems useful to also acknowledge this issue in the results or limitation sections to ensure readers are aware of it.

2) WS4: While I agree with WS that the t-test was not appropriate for your comparisons, I disagree with your choice to replace it with the Pearson chi-squared test. These are very different test that measure different things. While the former is a parametric test to examine differences in the means of an interval variable between two groups, the latter is used to examine differences between the expected and observed frequencies in a contingency table from categorical variables. However, your variables swimming competency, swimming distance, and swimming in the sea are all ordinal variables. Hence, a Wilcoxon rank-sum test would be able to indicate whether participants able to accurately identify the rip current were better or worse swimmers. This seems more in line with your original storyline and more insightful that just identifying differences in proportions without taking the ordinal character of the variable into account.

3) Presentation of statistical results (Section 3.1): I find your way of presenting proportions in the results section slightly confusion. For example, on line 157, you state “… higher proportions of males (33%, n = 15) that females (16%, n = 14) …“. It was initially unclear to me whether the 33% was 5 of 15 or 15 of 46. In some sentences, it is more clear what your reference class is (e.g., Line 189: “Of the 29 respondents able to spot the rip current, …”), but I recommend that you change your format to something like “(33%, 15 of 46)”, which avoid any misunderstandings.

4) Section 3.2: I find this paragraph quite confusing to read as you seem to be jumping around between different questions that you either interpret as interval or categorical variables. For example, you provide the descriptive statistics for the competence question on line 169, but you only mention that the difference is not statistically significant (which is fairly obvious) on line 174. I am also unclear on the reference class in your analysis of the confidence question. Did you include the responses from all participants or only the ones reporting that they were able to swim more than 100 m? If you used the entire sample, my calculations indicate that 32% among females (28 of 86) is significantly different from 54% among males (25 of 46) (chi-squared test: p = 0.025). I believe this paragraph would benefit from some clarifications.

5) Fig. 4 to 6: It is unclear to me why the format and styling of your stacked bar charts changes between Fig 4, 5 and 6. Fig. 4 shows % respondents by gender on the y-axis, in Fig. 5 it is number of respondents, and in Fig. 6 it is normalized percentages of respondents. In addition, the styling of Fig. 6 is different from the other two. However, as far as I can tell, there is no difference in how the three different graphs are interpreted. Hence, I recommend a consistent format and presentation of these charts.

I appreciate you attending to these challenges, which should not take much time. I am looking forward to seeing your fully revised manuscript.

Pascal Haegeli
NHESS Editor
Simon Fraser University, Vancouver, Canada