Scientifically, one should test multiple “identical” components for each test and average the results (or do something like a box and whisker plot) to incorpora

Scientifically, one should test multiple “identical” components for each test and average the results (or do something like a box and whisker plot) to incorporate manufacturing variability. But his results are authoritative as the bar is low.