Here's a proposal based on ideas gleaned from a recent discussion on this list:
Inputs are range ballots. For each alternative X tentative approval ballots are formed from the range ballots as follows: The alternative X is used as the approval cutoff. On some of the ballots the cutoff is considered excluded (not including X as approved) but on just enough ballots to make X the approval winner, the cutoff is considered inclusive (so on these ballots X is approved). The "reluctance" of ballot B in approving X is the difference between the maxrange value and the rating given to alternative X by ballot B. Elect the alternative X with the least possible reluctance total. In other words, find the alternative that requires the least total reluctant approval to become approval winner, and convert the range values to approval ballots using the minimum number of reluctant approvals of X to make X win. Then publish the approval totals, the total reluctance required to get X elected, and also for each of the other alternatives publish the voters' total reluctance at electing it. This method is summable. Unlike most DSV methods it is monotone. Does it satisfy consistency? Thanks, Forest ---- Election-Methods mailing list - see http://electorama.com/em for list info
