(Refer to the previous problem) The variance of the five replications for each run is calculated. Most of these variances are approximately equal but two are significantly lower than the others. The experimenters would be especially interested in those two runs if they want to optimize: