For a particular voter and a particular candidate, are their issue-differences on different issues additive? In other words, are a candidate's disutilities, for a particular voter, on various issues additive?
If yes, then city block distance is the more correct measure of distance.
If no, then Euclidean distance is the more correct measure of distance.
I suggest that, in principle, a candidate's various issue-disutilities for a particular voter can be expressed in terms of one kind of disutility, and added.
Money:
Placing a monetary value on all kinds of damages, including death, is routine in the court system.
So, if a candidate is going to kill lots of people in other countries, and is also going to cost you money with his fiscal policies, and is going to make you less safe with his war-on-drugs policies, these disutilities can all be put in terms money.
Loss of life-expectancy:
If a candidate's policies cause you to have less money, then his policies are reducing your choices, when making the right choice would maximize your life-expectancy. So his policies reduce your life-expectancy, assuming that you make good choices.
If his policies make the streets less safe, that reduces your life expectancy.
If his policies kill many people in other countries, those people have lost years of thieir life. Actual loss of years of life. Whatever their life expectancies were before the killilngs, they've lost that amount of life-expectancy.
A candidate's positions on polllution and job-safety can cause injuries and illness as well as death, but not only do those reduce life-expectancy, but the lower the quality of life. The notion of life-expectancy could be generalized to one that take into account the quality of life as well as its expected length.
It seems to me that life-expectancy is the better than money, since it's on a more solid basis to estimate the life-expectancy loss from monetary loss than to try to place a monetary value on loss of life.
But the point is that the various disutilities of a candidates, via various issue differences, can, in principle, all be put in terms of one undesirable thing.
So I suggest that city-block distance is the better measure of distances in issue-space.
With city-block distance, in spatial models, the CW always maximizes social utility.
But, as I mentioned before, the CW maximizes social utility with Euclidean distance, under the conditions that are usually assumed in spatial models.
Mike Ossipoff
_________________________________________________________________
Express yourself instantly with MSN Messenger! Download today - it's FREE! http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/
---- Election-methods mailing list - see http://electorama.com/em for list info
