I developed an IVRT test for a cream using 200 mL mini-paddle type 2 with immersion cell. Sometimes I observe one or two abnormally fast release rates, where most of the results are of a certain release rate with less than 10% RSD and the outlier results have a very different, faster release rates. The membrane is 0.22 um PVDF. I think the faster release rates may be due to thin sections in the membrane.

If the issue is the membrane then I feel it is wrong to use outlier results to determine release rate. Is it proper to reject these data? If so, how?