Prediction of Students’ Performances Using Course Analytics Data: A Case of Water Engineering Course at the University of South Australia
Round 1
Reviewer 1 Report
It must be renumbered in chapter 3. It is wrong where the numbering from chapter 2 was maintained in chapter 3.
I think in figure no. 4 you can include explanations about what means WA, AA, WA+.
In figure no. 3 (or in text) you can include percentage for all participants.
The article is good and this observations are minor.
Author Response
It must be renumbered in chapter 3. It is wrong where the numbering from chapter 2 was maintained in chapter 3.
Authors’ response: we have fixed the error of numbering in the manuscript.
I think in figure no. 4 you can include explanations about what means WA, AA, WA+.
Authors’ response: WA = Well above average and AA = Above average are included after Figure 4 in the revised manuscript.
In figure no. 3 (or in text) you can include percentage for all participants.
Authors’ response: The percentage of all participants are added in lines 113 – 115 in the revised manuscript.
The article is good and this observations are minor.
Authors’ response: Thanks for the comment.
Reviewer 2 Report
When using statistical test, as Mann-Whitney and Kruskal-Wallis in this case, the test statistic values and p-values must be reported. As to correlation analysis, it was not clear which coefficient was used but it looks like regression model with multiple regression coefficient R although the data was non-parametric. I suggest that you use Spearman’s rho which is a proper coefficient in this case.
as to parametric one-way ANOVA, there is no need to use it because non-parametric tests showed insignificant Neither it is not needed to tell all the basic things about the statistical tests used here.
As to the research ethics, it is expected to describe the process of getting informed consent from participants.
As to the conclusion a reader expects a more profound and inclusive discussion of the results.
Author Response
When using statistical test, as Mann-Whitney and Kruskal-Wallis in this case, the test statistic values and p-values must be reported.
Authors’ response: The test statistics and p values are added in lines 220 and 239 in the revised manuscript.
As to correlation analysis, it was not clear which coefficient was used but it looks like regression model with multiple regression coefficient R although the data was non-parametric.
I suggest that you use Spearman’s rho which is a proper coefficient in this case.
Authors’ response: The correlation analysis using Spearman’s rho was also performed and the value (0.22) is added in line 249.
As to parametric one-way ANOVA, there is no need to use it because non-parametric tests showed insignificant neither it is not needed to tell all the basic things about the statistical tests used here.
Authors’ response: We agree with the reviewer. However, the ANOVA analysis was performed only to validate the findings of the statistical tests, such as Mann-Whitney and Kruskal-Wallis. Section 3.4 shows the ANOVA test results.
As to the research ethics, it is expected to describe the process of getting informed consent from participants.
Authors’ response: We didn’t collect data using questionnaire survey, focus group discussion or interview process. So, a consent from participant may not be required. We collected data from course analytics. This research is approved by the ethics committee of the University of South Australia.
As to the conclusion a reader expects a more profound and inclusive discussion of the results.
Authors’ response: We added one more dot point in the conclusion section.
