My comment about the"pick-a-candidate" quiz ended with some thoughts on what I called"ranked preference" voting and the possibility of doing a national party primary poll instead of a dribble of early state votes. Anne Zook thought I was being silly, at least in my reliance on Internet technology. (Also check out the comments, in which we exchange some extended ranting about non-participants and I compare gerrymandering to treason. I am starting to get a little nervous about the percieved legitimacy of a system in which a minority of the population elects representatives who seem to work in the interests of their supportive constituents without a lot of attention to larger pictures. More on that some other time.)

She's right, in a sense: we haven't developed an on-line voting system that is as secure and reliable as our current paper ballot systems. I think that's because we haven't been trying very hard. When the stakes are high enough, it can be done: we send personal financial information over the Internet all the time, but most identity theft still relies on stealing data from merchants, public sources (how many colleges still use social security numbers for student identification?), or reconstructed physical evidence (receipts, etc.). But a national primary could be conducted in a number of other pretty secure ways. Mail-in ballots, for example, wouldn't be any less reliable than our current absentee ballots, and that system could certainly be tightened up with small changes. They are getting more popular, anyway, and have been shown to increase levels of participation (and Anne approves of them, too). Properly designed closed-circuit computer systems -- stand-alone networks without internet connections, like the networks which operate ATMs -- couldn't be hacked from the outside: terminals could be set up in voting districts, just like the primaries now. That's a couple of ideas.

But the method of voting is not as interesting to me, except as a technical issue, as the system of voting. The current dribble of early states is a classic case of evolution beyond function: there's no reason for 80+% of the voters to be excluded from the process because it's become a habit. If the technical aspects can be worked out, why don't we run a single national primary? We can make both primary day and election day holidays (someone recently suggested that we shift Veteran's Day to election day, which I like; we can use Memorial Day or Presidents' Day for primaries, depending on how long we want the general election to last.) and have some real fun.

If running a national poll with a single deadline seems too precipitous, there was a proposal floated four years ago to carry out"stacked" primaries, a kind of organized"Super Tuesdays" system. The states would be divided into groups (equal numbers in each group; I think ten was the original proposal), by size. The primaries would happen at one-week or two-week intervals in order of ascending size: the smallest states would vote one week (this preserves the value of voting in small states), followed by the next tier of states, ending up with a couple of weeks in which the bulk of delegates would be chosen. This has the virtue of being a relatively short process, compared to our current primaries, but one which distributes attention much more equitably."Momentum" would still be a factor, but it might actually mean something....

But my deep-rooted qualms about our electoral system have more to do with the actual voting process. One of the things I like about the Iowa caucus process is that it doesn't result in a winner-take-all delegate allocation for achieving a slight plurality. I'd be happy with either of two changes -- proportional allocation or Instant-Runoff Voting -- but both would be even better.

Instant-Runoff Voting (IRV), it turns out, is what I was calling ranked preference voting: voters choose multiple candidates, ranked first, second, etc. If there is no majority candidate, the candidate who recieved the lowest vote total would have their votes rescinded and those voters' votes would go to their respective second choices. This continues until there is a majority winner. As the current Z Magazine points out [by subscription only, but check out the Center for Voting and Democracy for lots more by the same authors] this will both strengthen alternative parties as well as reducing the"spoiler" effect of candidates like Ross Perot or Ralph Nader. Attack politics will be less attractive (if you can't be their first choice, you don't want to be the guy who called their first choice bad names) and coalition politics will rise quickly. If people don't have a second choice, they can just put a first choice. A little confusing? At first, perhaps, but people will get used to it pretty quickly, and just imagine what political scientists and pollsters will do with this data! At least it would reduce the kind of 'process' reportage that has turned our political process into a bad circus (not even a good circus, with clowns and trained animals; more like a bad county carnival with a few drunk jugglers.).

And let's just ditch the winner-take-all system of state votes, and give the top vote-getters (say, everyone over 10%) a proportional allocation of delegates: CVD calls this"Full Representation" or"Proportional Representation" (PR). So a candidate who was a strong regional candidate would still have to present good campaigns elsewhere because they would still be getting only a share of their states' candidates. On the other hand, they wouldn't be getting shut out of other states, either. It would slow down the selection process, and would also reduce the tendency towards attack campaigns (even if a candidate drops out, they still have delegates to broker). PR could also produce more parties: I foresee candidates appealing to different wings of their respective parties deciding to forego brokering deals and forming new parties.

Iowa already has PR, and a form of IRV because non-viable candidates' supporters get to pick a viable caucus to join. It can work (actually, it's fun, but I'm geeky). Before you say things like"well, it sure didn't reduce the attack campaigning in Iowa" remember that Iowa was just the first stop in a short string of winner-take-all winnowing exercises that may or may not last long enough for a majority of delegates to be chosen.

And all ballots should have"none of the above" as an option, and if"none of the above" wins they should hold the election over again, with different candidates. Seriously.