The measures are nearly identical for the two variables. I does not make much of a difference which variable is chosen. The Coefficient of Variation will be a good statistic.

CV=Standard deviation/mean

Dear  Dr.  Amin 

Thank you for this answer

i think that the Coefficient of Variation will be a good statistic if two variables have different mean but the above test have the same mean.

Regards

Abdel Rahim

I would use the average of both measurements, Abdel Rahim. 

Since the two datasets have the same mean and rather close SD, I would suggest calculating the SDs of the mean (SDM).     σm=σ/sqrt(n)    if sample sizes are nearly the same, then try to test the normality of both datasets.

It depends what metric you want to use, L sub 2, L sub 0, L sub infinity, or any other.  The decision which metric to use is entirely outside the mathematics you use and should depend on your knowledge of the problem and the goals you have.  If the values you show are coming from a statistical sample (random variable) then the two gentlemen before me may have some good points to consider.

Howgh.

The most appropriate approach is the use of coefficient of variation

CV =(Standard deviation)/(Mean) *100%