A couple of considerations ..
(a) do we see any hard evidence that a weighbridge system operating in real world conditions demonstrates statistically more reliable data than a well-controlled airline-style system ? Presumptions and good intentions are fine, but folk like Genghis and I, being anal engineers, are more likely to be swayed by numbers we can play with ...
(b) the comment re numbers of folk is valid for small aircraft except that the concern is population mass distributions rather than a ratio of passenger to aircraft weight. I suggest that the latter is a red herring argument.
However, there is a straight forward statistical approach to this problem, as has been done in Australia some years ago by CASA's John Klingberg. That study is not hard to get for reference.
Once we get to large aircraft, provided that the standard weight data is representative of the population on the aircraft (or vice versa, I guess) the standard weights work out fine.
I can remember numerous times having to weigh passengers .. only to find that the standard weight calc was so close to the scale result that we really had wasted our time ....