Updated: Maybe more relevant now. Posted previously with too much on the brain and came out wrong. Re-writing to be fair.:)
People here seem to vote in two classifications that I've seen. These are:
Technical Merit: In this classification, the ideals of a meritocracy hold true. A technical question is asked, and the reputation goes to those that show a flair for eloquence of coding, and obtaining the correct answer in the most efficient means.
Being as this is a very highly skilled group of individuals, then, in this forum, the -- and ++ can fly. Especially when slipups are made, and wrong answers given.
Philosophical: This is the one in which subjective opinion comes into play. There is a lot of personal bias at times, and sometimes, the --'s are given out on purely personal 'gut feelin', in much the same way that ++s are.
However, a lot of what I've noticed is that often --'s are given for bad argument stance. A higly polished argument, even if it's subjectively disagreed with often reaches a high ++ on the technical merits of it's proposing of an alternative stance.
This holds strongly to the tenets of debating, which, I was happy to find, are very alive and well here. :)

However, the anonymous voting is a good thing, as far as I can see. There are a large group of voters on hand, enough that the personal bias of one or two people simply get lost as 'line noise' in the general vote. A good post will still achieve a high reputation as the unbiased vote pushes it towards a realistic measure of it's worth.
Often, the -- of an article is a harrowing thing to receive. I believe I've had a fair few myself.. And honestly, I'd prefer not to know who did it.
The not knowing can often prevent such things as returning ++ for people who often rate up your articles, due to a feeling that you may owe a little, or --ing because someone has done that to you, colouring your view of the article in question, possibly then detracting from a clean rating.
Also, I don't believe a 1..10 scale is necessary.
The reason for this is that each of the monks has a limited amount of votes for the day. Thus, I try and take care to place those where they do the most good. I rarely --, unless something seriously crosses my ethics, or is a blatantly stupid technical post.
Most of the time it's better to leave a bad post to rot at rep. 0.
As the old saying goes, if there's one thing worse than being talked about, it's not being talked about.
Thus, a zero rep is about as shunned as you can get. :)
This limit of votes makes certain that someone has to have a high preference for that article in order to vote at all on it.
A 1 means that you got lucky.
A 10 is worthwhile.
A 20+ is very very interesting.
It achieves the same end with much greater simplicity.
A problem with a variable moderation is that for things to be fair, more 'votes' would have to be given. Otherwise, would you ever vote a '5' (average) with your vote? Most people would only vote on the lows (1 or 2 out of 10) or highs (9 or 10 out of 10), roughly approximating the ++ or -- already there.

As for my voting habits... Sometimes, I see code I recognise as being well thought out and having take a lot of time and effort to bring to us as a whole.. I'll ++ for the effort of someone doing that, even if it's not the best code in the world. I'll also ++ someone for going out of their way to be helpful, even if the enlightenment provided is only small, but in the right direction.
On the non-technical vote, it's either because I read something that makes me think long and hard about a viewpoint I've held for a long time (that gets ++ even if I still don't agree, usually), or if I see something I consider a well presented argument that says something I wish I'd written.. :)
I often have a few votes left at the end of the day, mainly because I don't have time to vote them all.. :)
HTH

Malk