This one has a reasonably settled answer, so here it is. Start from the measurement rather than the conclusion. Almost every disagreement here turns out to be two people measuring different things and comparing the numbers anyway.
That is the short version; the long version is somebody else's post.