Relative standard error (RSE)


I would like to do a comparison for two sets of data measuring the current (uA) in 25 points.

Sample 1
Mean current: 826.234
SD: 22.585
Sample size 25

Sample 2
Mean current: 797.426
SD: 17.026
sample size: 25

This will yield
Difference: -28.808
Standard error: 5.657
95% CI: -40.1816 to -17.4344
t-statistic: -5.093
DF: 48
Significance level: P < 0.0001

I would like to express the results as the relative difference of Sample 1 compared to Sample 2.
i.e. 28.808 / 797.426 = 0.0361 (3.6%), instead of the difference +/- the standard error

How do I calculate the relative standard error?
Intuitions says divide SE with the same reference value of 797.426 --> 0.00698 (0.7%).
It was difficult to find good information on how to use RSE in a simple comparison case like this.
Or are there better ways to go about this?
Thanks in advanced for any useful input!

Last edited:


Less is more. Stay pure. Stay poor.
Not sure of better ways. Could you just use a bootstrap. So get your above point difference estimate, then bootstrap the sample say a 1000 times and repeat the process for the point estimate.


I don't know why you can just divide the 95% CI values by 797.
Hi and thanks for your quick response hlsmith!
I have no experience in bootstrap statistics but the function is included in the Matlab statistical toolbox so I will at least give it a try and see what I'll end up with ;-)

Kind regards,