Bogus Balloting



The science of polling.

Loyalty: the tie that binds.

POSTED 16 NOV 2000 As Florida flails its way toward an electoral decision, the American voting system is under the microscope like never before. Were the ballots comprehensible? Do people count punch-card ballots better than machines? Sign reads: Polling Place, Vote here!, 6am to 7pm. Today. Should we deep-six the Electoral College and select the winner from the popular vote? Were the "chads" pregnant, for heaven's sake?

Unspoken, but critical in an election that probably was decided by Ralph Nader's third-party candidacy, is the question of dealing fairly with all candidates when more than two people are seriously running. (Remember Clinton, Bush senior and Ross Perot in 1992?)

Third-party candidates alter the electoral calculations of candidates and voters alike. If you're Al Gore, do you run toward Ralph Nader or toward George W. Bush? If you, as a voter, love Nader and tolerate Gore, should you vote for Nader, and hope the guy you really dislike -- W. -- still loses? Or should you put a clothespin on your shnozz and vote for Gore -- thus diminishing Nader's future clout?

Clothespin meets lesser evil
With third-party candidacies becoming more common, today's "plurality system" of voting forces voters to choose between "throwing away a vote" on a desirable but unelectable candidate and selecting the "lesser of two evils."

But plurality voting is not the only way to run an election.

So let the media moan and the politicians pontificate about the seemingly trivial problem of counting votes accurately. We're wondering -- if more than two candidates are running -- whether alternative voting systems would be fairer or more accurate than plurality voting .

Vote early, vote often
And cast your ballot for this lineup of alternative voting techniques.

Old illustration of Marquis de Condorcet (profile)Condorcet voting: Devised by an 18th century French mathematician named Marquis de Condorcet, this system asks voters to rank their preferences from top to bottom: say Bush over Nader over Gore. The results are used to compare candidates in pairs: this voter would prefer Bush to Nader, Bush to Gore, and Nader to Gore. If, as seems likely, Florida's Nader voters were closer to Gore than to Bush, Condorcet voting would have elected Gore. (The French may have had a remedy for voting woes - Condorcet lost his head in the revolution.)

What would happen in a polarized election where the three candidates were polling more equally? Say Mr. Lefty and Ms. Righty each pulled 40 percent of the vote, and Ms. Centrist got 20 percent. Under today's plurality system, the election would tilt on a (don't say it!) recount of the votes for Lefty and Righty (Centrist would have no more chance than Pat Buchanan in Palm Beach County). But with Condorcet, if Centrist was the number-two choice of partisans for Lefty and Righty, she would win.

Borda count: Named for a rival French mathematician (hey -- was Florida ever a French colony?) this system is similar but gives more weight to a person's preferred candidates, and less to the rest of the pack. Both methods can record subtler voter preferences than the plurality system.

Borda can be manipulated, however: Say a voter preferred Gore over Nader over Bush. If that voter thought Nader was more threatening to Gore, he or she might select Gore over Bush over Nader. By ranking the most serious rival to the preferred candidate last, this "strategic" vote sabotages Borda's purpose: identifying the candidate with the most support.

Cumulative voting: Each voter can distribute a certain number of votes among the candidates in any fashion. If the 2000 election allocated five votes per voter, a Gore-Nader straddler might devote three to Gore and two to Nader, while a Bush partisan could give all five to Bush. Cumulative voting benefits minorities, who can direct all their support to certain candidates, and it's been used to increase minority representation in Peoria, Ill., and Alamogordo, N.M. Like the Borda count, it can be manipulated by strategic voting.

Approval voting: Instead of ranking candidates, you "accept" however many you'd like, and the most accepted candidate wins. The system would allow the Gore-Nader straddler to approve of both candidates, without fretting about "throwing away a vote." While approval voting tends to reject minority candidates who might win a three-way plurality contest, minority candidates would at least benefit from a more accurate gauge of their support.

Given the potential advantages of alternative methods, is the United States stuck with plurality voting? Curiously, the U.S. Constitution has tons of detail about the Electoral College and four amendments on voting rights, but does not seem to specify the type of election. Apparently that remains a state decision.

A fly in the ointment
Before you get too hot and bothered about these candidate voting systems, we've gotta discuss a bit of math, and a rivalry dating back more than 200 years -- when Condorcet shot holes in the Borda method by proving it could fail. Condorcet thought his method was better -- but realized that it could likewise fail!

In 1952, mathematician-economist Kenneth Arrow examined voting systems that would meet some simple, logical premises:

If all voters prefer candidate Bland to candidate Boring, then Bland beats Boring,

The choice between Bland and Boring should not be affected (as it was in Florida) by other candidates, and

If the overall preference is Bland over Boring over Insipid, then Bland is preferred over Insipid.

The math is, well, a bit complicated for us simple voters at The Why Files, but since Arrow won a vote of the Nobel Prize Committee in 1972, we figure Arrow must have hit the bullseye.

His conclusion? The only voting system that satisfies all of these reasonable criteria is dictatorship! (Where is Fidel when we need him?)

University of Wisconsin-Madison mathematician Terrence Millar figures likewise. "In 1952, Kenneth Arrow showed that there can be no perfect voting system," says Millar. In fact, he says, more recent work by researchers such as H. Reiju Mihara prove that the theoretical problems with voting methods are even greater than Arrow understood at the time.

Yet despite the fact that any system can be subverted, misinterpreted or give mistaken results, Millar says, the Florida morass shows why "plurality voting is not necessarily the best way to assess voter preferences."

Had Florida voters been able to express more subtle choices, the outcome could have been quite different.

-- David Tenenbaum


Alternative Voting Systems.
Yahoo! scans the electoral systems.





Credits | Feedback | Search