There are several ways in which you can statistically address your problem.You can use either median or mode Just arrange the MIC values that you obtained in a serial order (low to high) and for an odd number of values, the middle value is the median, for an even number of values, the middle 2 values are averaged; the value occurring most number of times as MODE MIC. In both cases you can see if you are getting values that are comparable enough.

The mean. The mode gives you another information which is also very useful in terms of the frequency of occurrence and should assist in proper result presentation.

I have found out that dealing with agar dilution or disk diffusion methods, triplicate determinations are performed to validate the reproducibility of the test, taking the mean value as the MIC.

Thank you all for your help.

Is there any protocol where I can find the number of replicates needed? I mean, my impression (I'm not an expert in this field) is that microbiological protocols guide you through the experimental procedure, but not how to deal with the data collection. I agree with you that six replicates would be statistically stronger than three, but I can't index it because I haven't found bibliographic reference to support it.