DLC - Democratic Leadership Council
Democratic Leadership Council Home
Search Tips 

PrintPrintable Version of this Article

Send this Article to a FriendSend this Article to a Friend

File Attachments Exit_Poll_Stats


Related Links Politico: 'Study: Dems need white vote gain'

Press Release


Political Reform
The Vital Center

DLC | Political Study | September 25, 2008
Who Are the Swing Voters?
Key Groups That Decide National Elections

By Al From and Victoria Lynch

Executive Summary

  • Full report
  • Tables of exit poll statistics
  • If the history of recent general elections is a guide, the key to putting a Democrat back into the White House this fall will likely depend on how he fares with white voters with at least a high school education but no college degree.

    White men and women who have received their high school diplomas -- and those who graduated from high school and have attended some college while never getting a four-year degree -- have been critical swing voters in recent national elections. While Democrats rarely win a majority of them, those key voters vote significantly more Democratic in elections Democrats win than in elections they lose.

    Our estimate is that together white men and women with at least a high school education but no college degree swing the outcome of a general election by an astonishing average of 6.7 percentage points between elections that Democrats win and lose, respectively.

    That is more than double the margin by which President Bush defeated John Kerry in 2004. Cutting into Republicans' traditional margin with these voters could well mean the difference between a broad Democratic triumph and a narrow Democratic defeat.

    On average, white men with at least a high school diploma but no college degree swing the outcome 3.7 percentage points, and white women with the same education swing it 3 percentage points.

    To estimate the impact of swing voting among various categories of votes, we used exit poll results from six recent national elections. We first computed for each voter category the average marginal shift in Democratic voting between elections Democrats won and those they lost. Since the impact of shifting on the outcomes of elections depends on the relative size of the group, we then used the estimate of the categories relative size (among all voters in 2004) to calculate how many percentage points a Democratic candidate would win from the shift.

    For example, take white men with at least a high school education but no college degree. On average Democrats lost them in both winning and losing elections, but in losing elections they lost them by about 20 percentage points more than in winning elections. When that 20 percent shift is adjusted for size (white men with at least a high school education but no college degree made up 18 percent of all voters in 2004), the impact of their swing is 3.7 percentage points on the outcome of an election.

    A typical male voter in that category will likely be between 30 and 59 years old, live in a suburb or small town in the South or Midwest, and be married with no children living at home. He's likely to be a Republican or independent, moderate or conservative, not a member of a labor union, pro-life, and in favor smaller government. Finally, he's most likely to be Protestant but not a weekly churchgoer.

    His female counterpart has an only slightly different profile. She's also likely to be between the ages of 30 and 59, married with no children living at home, a Republican or independent, moderate or conservative, not a member of a union, pro-life, and for smaller government. She's most likely to live in a suburb in the South and have a gun in her household. Finally, she's more likely to be Catholic and a weekly churchgoer.

    Those conclusions and profiles are the principal findings of a study undertaken by the Democratic Leadership Council that analyzed exit poll data from the last five presidential elections and the 2006 Congressional election -- three national elections that the Democrats won and three that they lost. The purpose of this study was to identify voters who if recent historical patterns hold would most likely make the difference between a Democratic victory and defeat and who could be the key to a long-term Democratic majority.

    Among the other principal findings of our study were:

    • Despite all the talk about a rapidly changing electorate, there have been relatively small changes in the makeup of the voting electorate over the past 20 years, and the voting electorate in 2004 and 2006 remains remarkably similar to the electorate in 1988.
    • Certain categories of voters -- African- Americans, self-identified liberals, and voters who are strongly pro-choice -- voted overwhelmingly Democratic in every election regardless of which party won. Based on voting history, those three categories of voters constitute the Democratic Party's base.
    • About 40 percent of voters are part of the Democratic base (i.e., in one or more of the base categories in 2004). In 2004, John Kerry won about 80 percent of voters in those categories.
    • To get to a majority, Democrats must make up the difference by being competitive among categories of voters who swing back and forth between the two parties.
    • While African-American voters vote consistently and overwhelming Democratic, key segments of white voters tend to swing back and forth between the two parties. Overall, white voters are likely to swing the outcome of a national election by an average of 10 percentage points -- voting more Democratic in elections Democrats win and more Republican in elections Republicans win.
    • Two-thirds of that swing among white voters is accounted for by the white voters, profiled above, who have at least a high school education but no college degree.
    • To win a national election, Democrats don't need to win those key categories of swing voters, but they cannot be blown out among them. Losing them by 5 percentage points is likely to yield a Democratic victory; losing them by 25 percentage points is likely to yield a very different outcome.
    • In the 2004 election, base voters seemed more driven by issues and candidate characteristics that demonstrated compassion, while swing voters seemed more driven by issues and characteristics that showed toughness. To win over both groups and build an enduring majority, Democrats must demonstrate both the compassion to care and the toughness to govern.

    These findings are not intended to be predictive of the 2008 election, but rather to demonstrate how voters have cast their ballots in national elections. If recent patterns hold, the ability of Democrats to run competitively among key categories of swing voters could prove to be the difference between victory and defeat.

    Download the full report (PDF)
    Download tables of exit poll statistics (PDF)