Just a short section? Like how long? How many seconds above or below the level of another file was the measured file? I guess you see where I'm coming from...let's say you measured a single instance where one file was 1db louder than another file...and it was louder for a third of a second but every other second of the file duration was the same...then it's a stretch to say that there's much significant difference in volume.

I don't really care either way about whether x is different in volume than y, but I am curious what you actually measured.