What makes people confident that their testing method is yielding reliable results?
I guess it's rather subjective to a point. For my serious work and non-experimentation, I mainly do reversal and follow the box speed at first. If I personally don't find the rated speed to my liking, I assign it my own EI. But, I always make it a point to follow the manufacturer's specs for processing to the letter, this way I don't introduce any other variables that may cause inconsistencies.
"The real work was thinking, just thinking." - Charles Chaplin