Party Parity
ABCNEWS Poll Finds Country Evenly Split Between Dems, GOP
abcnews.go.com
Analysis
By David Morris, Telis Demos and Gary Langer
N E W Y O R K, Nov. 4— A year from the next presidential election, the nation stands at a rare point of political parity: Across 2003 precisely equal numbers of Americans have identified themselves as Democrats, Republicans and independents, a first in 23 years of ABCNEWS polling.
The year's averages — 31 percent for each group — mark an uneven but long-term rise among Republicans, to a new high, and the fewest Democrats in annual averages since 1981. All else being equal, the trends suggest continued Republican competitiveness in election politics, albeit far from the Democrats' onetime dominance in sub-presidential races.
The vagaries of politics can change the scene. Rather than straight lines, party ID has played out in a long series of forays and retreats over the years, as the parties have battled chiefly over a small group of lightly committed Americans, mostly independents.
Change has been so gradual that, projecting past trends into the future (never a sure bet), it would take another 10 or 12 years for Republicans consistently to outnumber Democrats.
But time has not been on the Democrats' side. There is a significant relationship between the annual average number of Democrats versus Republicans and the passage of time since 1981 (a correlation of .71, where 1 is strongest, 0 weakest).
This analysis is based on a review of 332 ABCNEWS and ABCNEWS/Washington Post polls, comprising more than 350,000 interviews conducted from 1981 to present. This year's sample alone includes 22 surveys with 1,000 or more interviews each. The size of the dataset means even small changes are statistically significant.
Factors
Two immediate factors appear to have helped push the Republicans to parity: First was winning the White House in 2000; the number of self-identified Republicans advanced from a 27 percent average in the 2000 election year to an average of 30 percent during President Bush's first year in office.
That's not the norm. It did happen after Ronald Reagan won re-election in 1984, with Republican self-identification going from 24 percent that year to 28 percent in 1985. But GOP partisanship didn't advance after George H.W. Bush was elected in 1988, nor did Democratic allegiance in 1993 or 1997, after Bill Clinton's two elections.
The second bump came from a rally behind George W. Bush after the terrorist attacks of Sept. 11, 2001. (His job approval rating soared to a record 92 percent a month later, and remained above 66 percent for more than a year.) Republican self-identification came within two points of the Democrats' in 2002, and then went on to match it this year.
Party ID is determined by more than individual personalities and events; it's also tied to basic political philosophy. People who prefer a larger government that offers more services tend to be Democrats; those who'd rather see a smaller government that does less are more apt to be Republicans. When, on Jan 27, 1996, Clinton declared, "The era of big government is over," he was reflecting a philosophical shift that more naturally fits Republican views.
Independents and Others
Contrary to conventional wisdom, party ID data do not show a long-term rise in the number of independents (as those who posit widespread weariness with the major parties would suggest). Instead, at 31 percent in 2003, independents are at about the same level as their long-term average, 32 percent since 1981.
One change has been on the edges: Third-party identifiers, people who are neither Democrats, Republicans nor independents, have made up six percent the last two years and seven percent in 2001 — triple their level in the late 1980s and early 1990s and two points over their long-term average.
Parity, evidenced by the all-but-tied 2000 contest and Ralph Nader's role in denying Al Gore a victory, gives this small group clout.
Also telling is "leaned" party identification, which adds to the tally unaligned Americans who lean toward one of the two parties. Here Democrats continue to outnumber Republicans, but their edge — three points, 46 percent to 43 percent — is far narrower than it's been. A decade ago (during Clinton's presidency) it was a 13-point lead, 51-38 percent — the last year in which the Democrats could claim majority status. And the 1981-2003 average is 49-41 percent.
Vote
While attracting more adherents can't hurt a party's performance at the polls, it doesn't guarantee victories. One reason is that party ID informs vote choices, but doesn't dictate them; a popular candidate or a pressing issue — such as the economy, war or scandal — can move the middle far faster than can gradual shifts in party allegiance. This is particularly true in presidential races, which tend to be more personality-driven.
Also, in election politics, what matters more than partisanship across the population is turnout on Election Day. Republicans long have been more likely than Democrats to vote; they've come within three or four points of Democratic party ID among voters in every presidential election since 1988 — closer than party ID in the general population, including nonvoters.
One question now — for 2004 and beyond — is whether Republican parity in the broader population is reflected in voter turnout.
Another problem for the Democrats is their harder time maintaining party loyalty at the polls. In 2000, Bush held 91 percent of Republicans, while Gore won a smaller share of Democrats — 86 percent. Clinton held his base slightly better than did Bob Dole in 1996 or George Bush in 1992, and in '96 Clinton won 13 percent of Republicans.
But in 1988, 17 percent of Democrats bolted to Bush's father rather than support Michael Dukakis. And a quarter of Democratic voters abandoned their party in 1980 and 1984 alike — the famous "Reagan Democrats."
Whatever happens at the presidential level, the change over time in party ID has been accompanied by a change in the division of the nation's other political spoils.
The Democratic Party controlled 23 more state legislatures than did the Republicans in 1983, and 24 more in 1990; today, Republicans control five more. There were 18 more Democratic than Republican governors in 1983; today, there are two more Republican governors, not counting the yet-to-be sworn in Arnold Schwarzenegger (although the GOP advantage in governorships was higher, +15, in 1997). And in 1983, the U.S. House had 103 more Democrats; today it has 24 more Republicans.
There are strong correlations between the passage of time since 1981 and party holdings of state legislatures, governorships and House seats: .75, .72 and .87 respectively (again, where 1 is strongest, 0 weakest).
Timeline
In the period reviewed here, Democrats peaked at 39 percent of the population on average in 1983, slipped to 35 percent by 1993, and dipped further to this year's 31 percent — three points below their 1981-2003 average.
Republicans moved in the same period from 23 percent in 1983 to 26 percent in 1993 and 31 percent today; they're now three points over their average for the period. Independents, as noted, are very close to their average (a point below it).
Independents are more changeable in the short-term; while relatively stable over time, their numbers tend to fluctuate more than either Democrats' or Republicans'. Standard deviations — a measure of variation from the average — are about the same for Democrats and Republicans (2.8 and 2.9 respectively), but higher for independents, 3.5.
Illustrated another way, in 2003, independents have ranged from 27 percent of the population — the lowest among all three groups — to 37 percent, the highest.
All three groups have gone through broad ranges in support; in these 332 surveys Democrats have ranged from a high of 43 percent to a low of 28 percent; Republicans, from a high of 34 percent to a low of 19 percent; and independents, from 40 to 23 percent.
Short-term events can influence these individual readings, as can normal sampling variability (although to minimize it, only surveys of 1000 or more interviews were used in this analysis). Using annual averages smoothes the results to clarify long-term trends.
Unleaned Party ID
1983 average
Democrats: 39%
Republicans: 23%
Independents: 34%
1993 average
Democrats: 35%
Republicans: 26%
Independents: 35%
2003 average
Democrats: 31%
Republicans: 31%
Independents: 31%
1981-2003 average
Democrats: 34%
Republicans: 28%
Independents: 32%
Ebb and Flow
As noted, the road to parity is littered with successes and setbacks for each party. Small, often short-term shifts in party ID can follow presidential rallies and slumps — up for Reagan and the GOP, for example, in a rally after the Marine barracks bombing and the invasion of Grenada in 1983; down during the Iran-Contra scandal.
Overall, rebuilding from a long, Watergate-induced slump in the 1970s, the Republicans gradually gained ground through the 1980s, peaking during the 1991 Persian Gulf war (33 percent Republicans, 32 percent Democrats in consecutive February 1991 surveys). But postwar economic discontent led to a resurgence for the Democrats, and George Bush's loss to Clinton in 1992.
Clinton, on his watch, took office with an eight-point advantage for Democrats, only to see Republicans close the gap after dissatisfaction with his performance, continued economic discomfort and the "Contract with America" campaign gave them control of the House and the Senate in the 1994 midterm elections.
The GOP again was well-positioned. But the economy — and Clinton — rebounded, Gingrich faltered, and the Republicans lost steam. Clinton won a second term, and the public went on to side against his Republican-led impeachment. In an ABC/Post poll immediately after the House voted to impeach Clinton, 35 percent of Americans identified themselves as Democrats, just 22 percent as Republicans.
Leaners
Independents are the quintessential swing voters; the candidate they favor on Election Day invariably wins. (2000 was anomalous; Bush had a slight edge among independents, 47-45 percent, but narrowly lost the popular vote.) So where independents lean matters much.
Leaned party ID in 2003, 46 percent Democrat to 43 percent Republican, is much like the 2002 average (47-44 percent) but otherwise a departure. It was 49-41 percent in 2001 and quite similar back to 1996. Leaned party ID has narrowed before — in 1985 and '86 (Reagan years), 1990 and '91 (Bush/Iraq war years), and 1995 (after the GOP took Congress.)
One reason for the fluctuation is that in politics, as in dating, some people would rather play the field than make a long-term commitment; so it appears to be, particularly with a fairly small group of lightly committed independents. This year's war in Iraq underscores this behavior, with shifting allegiances of a relatively few people causing poll-to-poll fluctuation among leaners as the long-term numbers stay fairly stable.
In a February poll, just before the war with started, leaned Democrats outnumbered Republicans by 48-43 percent. But at the end of April, as major fighting wound down, each party registered 46 percent. Then it was back to a seven-point Democratic advantage among leaners in a July survey as the postwar effort dragged on and casualties mounted.
But even a big advantage among leaners doesn't guarantee election victory. Clinton and Walter Mondale went into elections against incumbent Republican presidents with the same numbers in leaned party identification, 51-39 percent. In 1984, Mondale lost to Reagan by 18 points. Eight years later, Clinton defeated George Bush by 43-37 percent.
Ideology
Conservatives account for about a third of Americans, liberals about a fifth, and moderates just over four in 10. That gives Republicans another advantage &30151; a bigger base from which to draw.
And they are drawing from it more broadly: Today nearly half of conservatives are Republicans, compared to just over one in three in 1983.
Republicans also have held a steady share of moderates. The Democrats have held stable — but not grown — in their liberal base, and there are fewer moderate or conservative Democrats than previously.
Other Groups
Age also has been working for the Republicans; they've improved among young adults, while Democrats have held their share of older Americans.
About a third of adults under age 30 now say they're Republicans, up from about a quarter in 1983. Meanwhile the Democrats' share of young adults has gone from about a third 20 years ago to fewer than a quarter today. (Among older adults, four in 10 are Democrats.)
In addition to being older, Democrats these days are more apt than the broader population to be women, nonwhites, and at the lower half of the income scale. About a quarter of whites are Democrats, compared with more than a third each who are Republicans or independents. Majorities of blacks and people of other races are Democrats.
About a third of women are Democrats, compared with about a quarter of men. About three in 10 women and men alike are Republicans, while among independents, men slightly outnumber women.
Methodology
This analysis is based on 332 ABCNEWS and ABCNEWS/Washington Post polls from 1981 to 2003, each composed of 1,000 or more random-sample telephone interviews. Respondents were asked, "Generally consider yourself to be a (Democrat) a (Republican), an independent, or what?" In a subset of these polls, respondents who didn't choose one of the major parties were asked: "Do you lean more towards the Democratic Party or the Republican Party?" |
