With another close presidential contest in store, that hardy if indecipherable oddity of American politics, the electoral college, is back in the news. My esteemed colleague William Raspberry has lent his powerful voice to those who for decades have railed against the injustice of the unit rule, which gives all of a state's electoral votes to the candidate who wins a plurality from its voters.

Because of that rule, Raspberry wrote in a column the other day, nearly half of the Floridians who battled butterfly ballots and official obstacles to vote in 2000 "might as well have stayed at home," because a tiny margin of 537 votes in the official tally gave George Bush all 25 of the state's electoral votes -- and the presidency.

The same complaint could be made by those who voted Republican in Iowa, Wisconsin or New Mexico, narrowly won by Al Gore. The unit rule is used everywhere except in Maine and Nebraska, which award one electoral vote to the winner of each congressional district.

To solve the problem that vexes Raspberry and a great many others, a Colorado voter initiative on next month's ballot would divide the state's nine electoral votes according to the share of the popular vote each candidate wins. That is only one of the proposed remedies that have been considered -- and one of the easiest to debunk. If the proportional system, as it is known, became the standard for all states, the most predictable effect would be to throw more presidential elections into the House of Representatives.

A study by Congressional Quarterly, quoted by professor Judith Best of the State University of New York at Cortland in the Spring 2004 issue of Political Science Quarterly, found that at least four of the elections since 1960 -- those in 1960, 1968, 1992 and 1996 -- would have gone to the House under that system. The 2000 election might have wound up there, too, depending on how fractional votes were rounded. The Constitution requires someone to win a majority of electoral votes; otherwise, the House chooses the president from among the top three finishers.

How do you think the public would react to the discovery that in such a contingent election, each state delegation has one vote, regardless of its size -- the Democratic majority from California being matched by the single Republican member from Delaware?

Because that idea seems so flawed, most of those who support electoral-college reform favor going all the way to direct national election of the president. A constitutional amendment to abolish the electoral college and substitute direct election actually passed the House in 1969, only to fail in 1970 and again almost a decade later in the Senate.

In the same issue of Political Science Quarterly, professor Jack Rakove of Stanford makes the modern case for direct election. He points out that it would force candidates to compete everywhere, including in such "safe" states as Texas and New York, whose residents now see them only at fundraisers.

Modern polling, he correctly notes, allows candidates to target their appearances and their advertising on a few closely contested states -- fewer than one-third of the 50 at any point in this campaign -- and virtually ignore the rest.

Direct election would also end what he calls the weighting of the electoral-college system toward small-population states because each state, no matter how lightly populated, is awarded a "bonus" of two electoral votes.

But direct election, however appealing, has plenty of problems built into it. When Congress debated it after George Wallace threatened electoral deadlock with his third-party candidacy in 1968, opposition came from small states, whose senators feared they would be overlooked by the candidates, and from urban constituencies, who feared diminution of their power to swing big blocs of electoral votes through the unit rule.

A bigger problem, Best and others argue, could be the effect on the two-party system. Most proposals for direct election specify a minimum percentage for victory -- usually 40 percent or 45 percent -- with a runoff between the top two contenders if no one reaches that threshold.

But as soon as you introduce the possibility of a runoff, you create an incentive for minor parties to form, in hopes of bargaining for favors or policy concessions from the runoff opponents. In such a system, a John McCain might have continued running after the primaries of 2000 to extract a promise from Bush to sign campaign-finance reform, or a Howard Dean this year in hopes of swaying John Kerry's policy on Iraq.

I suspect this whole electoral-college issue is due for serious debate in the next Congress. But prudence dictates a long, skeptical look at the seemingly easy solutions.