Many comments about approval strategy have made little sense to me, because they ignore the fact that strong supporters of a particular candidate can have very divergent views about other candidates. Using the 2004 U.S. presidential election as an example, supporters each candidate, from the socialist candidates on the "left" to the Constitution Party candidate on the "right," came in many different varieties, depending on what their particular sets of major concerns were and the strengths of their various concerns. (I've put the words left and right in quotes because I don't think they are as usefully descriptive as most people assume they are.)
For example, many people who own significant amounts of stock might argue that only Bush and Kerry were acceptable choices because the election of any other candidate could have caused great economic uncertainty and endangered the value of their stock holdings. Others who really liked one of the third party candidates might have ranked Kerry and Bush a close second and third (or vice versa) and approved all three of them in an approval election because while they liked what the third party candidate was proposing, they also liked the apparent economic stabililty of electing one of the major party candidates. Then there are people who were much more strongly opposed to either Kerry or Bush than they were in favor of any other candidate. In fact, there appeared to be lots of "anybody but Bush" and "anybody but Kerrry" voters. In the case of such voters, the best strategy in an approval election would be to approve every candidate who appears to have any chance at all of defeating the candidate they are particularly strongly opposed to. Given the great diversity of voter opinions on a variety of key concerns, I just don't see how it is possible to state formulas that are optimal for all or even most supporters of particular candidates. -Ralph Suter ---- Election-methods mailing list - see http://electorama.com/em for list info