In every presidential election since 1980, men have been more Republican than women, but they have both voted for the winner. Could 1996 be the first year in which the gender gap is a decisive factor?

The special Senate election in Oregon in January was a breakthrough, but not for the obvious reasons. Not because Democrats took a seat away from the Republicans for the first time since Bill Clinton was elected. Not because it was the first statewide election for federal office conducted entirely by mail.

It was a breakthrough because it brought gender politics in the United States to a new level. Just beneath the surface of a very close race (Democratic Rep. Ron Wyden defeated Republican Gordon Smith by 1 percentage point) there was an unusually sharp split between men and women. And an unusually heavy mobilization of women voters. The Oregon results could presage a Democratic resurgence this year on the strength of what the early-February Feminist Expo 96 called "women's empowerment."

The gender gap has been a feature of all recent presidential elections, but it has never been decisive. The last pre-gender gap election, in 1976, was close, but men voted exactly the same as women, according to the network exit polls: Men and women alike voted for Democrat Jimmy Carter over President Ford, 50-48 percent.

The gender gap sprang up with Ronald Reagan's candidacy in 1980. Men voted 8 percentage points more Republican than women did. Nevertheless, both men and women voted for Reagan over Carter--men by a 19-point margin, women by just 2 points. The same thing happened in 1984. Men voted for Reagan over Walter F. Mondale by a whopping 25 points. Women favored Reagan by 12 points.

In 1988, men went for Republican George Bush over Democrat Michael S. Dukakis by 16 points. Women voted for Bush, too--by 1 percentage point. Women voted for Clinton over Bush in 1992, by 9 points. But men also voted for Clinton, by 3 points.

In every election since 1980, men have been more Republican than women. But they've both voted for the winner.

That didn't happen in the Oregon special election. It was a close election, but men and women were miles apart. That's what Voter News Service discovered in a poll of Oregon voters on behalf of the major television networks and the Associated Press. The poll found that men voted for Smith by a 10-point margin. Women voted for Wyden by 7 points. So how did Wyden win?

The answer: 57 percent of Oregon voters were women. That's an extraordinary turnout by women. Women were 50 percent of Oregon voters in the 1992 presidential election and 51 percent in the 1994 midterm balloting.

What drew all those women to the polls? The key factor seems to have been a backlash against the Republican Congress.

Supporters of the two candidates had very different agendas. Those who voted for Smith named the federal budget and taxes as their top issues. Those happen to be the issues at the top of the Republican agenda in Congress as well. Among those who voted for Wyden, education and the environment were the biggest concerns. In Clinton's recent State of the Union address, he named education and the environment, as well as medicare and medicaid, as the four programs he intends to protect from Republican cuts. The concerns of Oregon voters reflected the national concerns of the two parties.

Smaller numbers of Oregon voters said that they voted explicitly to send a message to Washington. Those who said that their major concern was Republican policies in Congress voted strongly for Wyden. Those who said they were concerned about Clinton's performance as president voted strongly for Smith. Twice as many Oregon voters were concerned about the Republicans (12 percent) than about the president (6 percent).

It was a reversal of 1994. In 1994, voters all over the country rushed to the polls to register their anger at Clinton. In Oregon last month, voters angry at Clinton were outnumbered by voters angry with the Republican Congress. That was enough to account for the margin of difference in the outcome.

Wyden owes his election to the women of Oregon. They led the backlash against the Republican Congress, just as angry white men led the backlash against Clinton in 1994.

Could the gender gap save Clinton? In 1992, he carried Oregon by 10 percentage points. In last month's Voter News Service poll, he was running 15 points ahead of Robert Dole in Oregon. Among men, the Clinton-Dole race was a dead heat. Among women, Clinton had a 26-point advantage.

The Oregon election was unusual in that it was conducted entirely by mail. Did that have anything to do with Wyden's narrow victory?

Holding an election by mail creates some problems. There is no guarantee of a secret ballot. Voters can mail in their ballots long before the campaign has ended. Campaigns have to spend a lot of money targeting people who have not yet voted. And it eliminates one of the few collective rituals in American civic life: Election Day, where an entire community joins together and comes to judgment.

Oregon voters, however, thought the mail ballot was just fine. Twice as many (56 percent) said they preferred a mail ballot to voting in a polling place (27 percent). Turnout was unusually high. Two-thirds of Oregon's registered voters cast their ballots by mail. That's a lot higher than the 38 percent who voted in 1994. It's even higher than the 50 percent who voted in the 1992 presidential election.

Does Wyden owe his victory to voters who mailed in their ballots but would not have bothered to go to the polls? A fourth of those who voted in the Oregon special election said they didn't vote in 1994. If they didn't vote in the midterm elections, they would have been unlikely to vote in a special election had it not been for the convenience of a postal ballot.

Among those who voted by mail last month but did not vote in 1994, Smith narrowly led Wyden. That disproves the theory that the additional voters brought out by the postal ballot were the key to Wyden's victory. They were not. Women were.

William Schneider is resident fellow at the American Enterprise Institute.