Most Americans said that their decision in Tuesday's stunning election was not a vote for or against President Clinton.
But like a subterranean river, attitudes toward Clinton, the Republican Congress and the bitter battle over impeachment all helped to propel the Democrats' modest yet historic gains, according to a national exit poll conducted after the balloting.
Though the electorate was slightly less positive toward Clinton than the country as a whole, almost two-thirds of those voting said Congress should not impeach the president--and those voters preferred Democrats over Republicans in their congressional districts by more than 2 to 1, according to the exit poll. Likewise, just over three-fifths of voters said they disapprove of the way Congress has handled the Clinton controversy--and they gave fully 68% of their votes to Democrats.
"It's pretty clear," argued Democratic pollster Mark Mellman, "that people do not want to go forward on impeachment."
The poll found that, although most voters expressed unfavorable opinions of Clinton as a person, they continued to give him much higher marks for his job performance than for Congress'--and that advantage also helped Democrats, as they became the first party holding the White House to gain seats in a midterm election since 1934.
In reversing that historical pattern, Democrats benefited from a strong turnout among their core supporters--including African Americans and union members--and a better showing among men and both middle- and upper-income voters than they managed while being swept from control of Congress in 1994.
Meanwhile, religious conservatives, who had been expected to storm the polls in protest over Clinton's affair with Monica S. Lewinsky, constituted a slightly smaller share of the electorate than in 1994.
Common Patterns in Local Results
Midterm elections are a collage of local contests, each shaded with its own distinctive colors. Yet the election exit polls--conducted for The Times by the Voter News Service, a consortium of newspapers and television networks--illuminate common patterns in the diverse local results.
This year's poll--which surveyed 10,017 voters as they left 250 precincts across the nation--shows how small changes in the electorate can produce enormous changes in results when the country is as closely divided between the two parties as America is today. Overall, the two parties split the congressional vote almost exactly in half, the survey found.
The polls registered only small changes in the composition and preferences of the electorate since the first midterm election of Clinton's presidency in 1994. But those were enough to make the difference between the Republican landslide four years ago and the Democratic showing Tuesday that reduced the Republicans to the narrowest House majority since 1931 and denied the GOP any new seats in the Senate.
Compared to 1994, Democrats improved their performance among men by 4 percentage points, and by about the same amount among voters earning from $15,000 to $50,000 a year. Likewise, compared to 1994, Democrats gained by 3 percentage points among high school grads and by 5 percentage points among voters with some college but not a four-year degree--working-class groups that abandoned the party in the GOP sweep.
But Ruy Teixeira, an opinion analyst at the liberal Economic Policy Institute, pointed out that Democrats still struggled with those groups in 1998--only leading with high school graduates by 2 percentage points and narrowly losing voters with some college experience. In fact, he pointed out, the Democrats actually lost ground with both groups compared to 1996, when Clinton's reelection helped the party regain nine seats in the House.
"If you are looking for a Democratic strategy that will rebuild a majority with white voters of moderate income, clearly they are not there yet," he said.
If anything, the exit polls showed, compared to both 1996 and 1994, Democrats enjoyed their largest gains among voters earning $75,000 a year or more--a group that apparently gave Clinton and his party credit for good economic times.
Democrats benefited not only from changes in preferences but from shifts in the composition of the electorate. Union households, which gave three-fifths of their votes to Democrats, constituted 22% of the electorate, about the same percentage as in the presidential year of 1996 and significantly more than the 14% that turned out in 1994. African Americans, who preferred Democrats over Republicans by 8 to 1, ticked up slightly from 9% in 1994 to 10% this year. Latinos, who gave Democrats almost three-fifths of their votes, also edged up from 3% in 1994 to 5% this year.
By contrast, religious conservatives, who constituted 15% of the vote in 1994 and 17% in 1996, surprisingly dropped to 13% this year. That falloff was part of a broader decline in the share of the electorate that described itself as conservative--one of the key trends in the voting.
Liberals, conservatives and moderates this year all split their votes between the parties in almost exactly the same proportions as they did in 1994. But, continuing a trend that appeared in 1996, the share of the electorate that defined itself as conservative dropped (from 37% in 1994 to 34% in 1996 to 31% this year), while the slice that considers itself moderate increased, to half of the total vote. Since Democrats enjoyed an 11 percentage point advantage among moderates (while trailing by 63 points among conservatives) that change explained much of their gain on Tuesday.
Sorting out the impact of the impeachment controversy on the election remains a complex task. On the most direct question, few voters were inclined to view the election as a referendum on impeachment.
Most Voters Weren't Sending a Message
Asked whether they were voting to express support or opposition to Clinton, 60% of voters said neither. The rest split narrowly between those who wanted to express opposition (21%), and support (18%) for the president. Similarly, voters divided in a virtual three-way split when asked whether they were voting to express support for Republican control of Congress, opposition to it or neither.
Yet the scandal plainly motivated some core Republican voters: 18% of those surveyed said moral and ethical standards was the principal issue influencing their vote and they gave Republicans four-fifths of their ballots. And voters who opposed impeachment--the clear majority of the electorate--provided the decisive margins for Democrats.
On the bottom-line issues, 63% of those voting said Congress should not impeach Clinton, 56% said he should not resign and nearly three-fifths said Congress should "drop the whole matter" without conducting hearings. Democrats carried comfortable majorities in each of those groups.
Those numbers cannot answer whether those voters supported Democrats because they opposed impeachment. But there is no question that attitudes not only on impeachment but on the president's performance in office all closely tracked the congressional vote. This year, three-fourths of those who thought Clinton was doing a good job voted Democratic for Congress--about the same percentage as in 1994. The big difference: in 1994, just 44% of voters approved of Clinton's performance. On Tuesday, 55% did.
Six weeks ago, Congress' approval rating was also running at 55%. But it slipped as Congress authorized the impeachment inquiry, and just 41% of voters gave Congress positive marks.
That decline proved to be one of the keys in the election. Although those who approved of Congress' performance gave two-thirds of their votes to Republicans, Democrats won support from three-fifths of those who disapproved--and at 55% of the electorate, that was the much larger group.