How to round off percentage uncertainties?

[jayadds]

How to round off percentage uncertainties?

Here's is an example question:

Find the percentage of uncertainty in (225.7+/-0.5)mm.
I know that to find out percentage, it's simply 0.5/225.7 x 100 = 0.22153088...%

Now, my question is, how much do I round off for this percentage uncertainty? Should I just round it off to one significant figure? Or Is there a certain rule?

 

[technician]

strictly speaking you cannot use more sig figs than the smallest you are given/calculate...that is what the word 'significant' means (If they are not significant then they are insignificant !)
Your error is quoted as +/-0.5mm which means 1 significant figure so I would say round off to 0.2%
I would challenge anyone dealing with error analysis to justify that this should be given to 2 or 3 sig figs.