Reviewing statistical analysis and fixing the issues identified in a manuscript
Budget: $30 – $250 USD
Hello,
I had the statistical analysis for the data collected from an educational survey done by a data analyst, but the reviewer of my research identified gaps in the data. Below is a summary of the feedback received. If you are able to help fix these issues, please let me know.
Thanks
3) In the data analysis section, the SEM approach was mentioned, but needs justification. Also, you need to mention the software SmartPLS which you used to conducted the SEM-PLS analysis.
4) As I can see high correlations among variables (Table 2), it would be better to examine the discriminant validity. Although it is mentioned, I cannot see it. And same for the convergent validity estimates. Both are doable in SmartPLS software.
5) I like the cluster analysis, but I got limited knowledge about it.
6) The hypotheses are not found
7) I think the second-order construct analysis (as shown in Figure 5) is not done correctly. I expect to see either all items associated with each construct, or to set a second-order constructs where the first order constructs are not shown as items (yellow rectangles). I actually cannot understand whether the scores between the items (rectangles) and the constructs (circles) are beta values or item loading. Those scores might be extracted from the T-statistics (bootstrapping), not the original PLS algorithm. Also, the scores of negative motivation and engagement should be negative. I suggest to fit it.
8) The same issue here. The results in Table 6 contradict the results in Figure 5. The moderating effects are not found significant based on the table, whereas the results of the moderating effects in Figure 5 show plausible scores. The same problem in point 7.
9) Figure 6 looks better (in first-order construct). The signs among negative constructs and student achievement should be negative, though.
10) Based on the P-values, 6 hypotheses were not found significant. Hope you can justify this properly.
I had the statistical analysis for the data collected from an educational survey done by a data analyst, but the reviewer of my research identified gaps in the data. Below is a summary of the feedback received. If you are able to help fix these issues, please let me know.
Thanks
3) In the data analysis section, the SEM approach was mentioned, but needs justification. Also, you need to mention the software SmartPLS which you used to conducted the SEM-PLS analysis.
4) As I can see high correlations among variables (Table 2), it would be better to examine the discriminant validity. Although it is mentioned, I cannot see it. And same for the convergent validity estimates. Both are doable in SmartPLS software.
5) I like the cluster analysis, but I got limited knowledge about it.
6) The hypotheses are not found
7) I think the second-order construct analysis (as shown in Figure 5) is not done correctly. I expect to see either all items associated with each construct, or to set a second-order constructs where the first order constructs are not shown as items (yellow rectangles). I actually cannot understand whether the scores between the items (rectangles) and the constructs (circles) are beta values or item loading. Those scores might be extracted from the T-statistics (bootstrapping), not the original PLS algorithm. Also, the scores of negative motivation and engagement should be negative. I suggest to fit it.
8) The same issue here. The results in Table 6 contradict the results in Figure 5. The moderating effects are not found significant based on the table, whereas the results of the moderating effects in Figure 5 show plausible scores. The same problem in point 7.
9) Figure 6 looks better (in first-order construct). The signs among negative constructs and student achievement should be negative, though.
10) Based on the P-values, 6 hypotheses were not found significant. Hope you can justify this properly.