Hello Dr Kp

The patients are independent to each other and you are not following up the same patients at the different time periods to measure their characteristics that's why the observations and groups are independent therefore independent t-test must be used to compare the means of these characteristics.

Thank You.

Thank you for your attention, Mr Husain.

The approach you have described is the one I would have chosen.

However, I recently came across to this article:

https://www.ahajournals.org/doi/pdf/10.1161/HYPERTENSIONAHA.107.099986

Here the authors have divided a group of patients into two subgroups based on endothelial function. To compare these two subgroups, they employed Wilcoxon signed-rank test which is a non-parametric test for two related samples.

Could you/or anyone else please comment on the rationale behind their choice?

Thanks,

KP