PRINCETON, NJ -- If the elections for Congress were held today, 50% of U.S. registered voters say they would vote for the Democratic candidate in their district and 44% for the Republican candidate.
Â鶹´«Ã½AV uses this "generic" ballot to measure support for the two parties in congressional elections. Historically, the final pre-election generic congressional ballot based on likely voters has proven to be an accurate predictor of the national two-party vote for the U.S. House in past midterm elections. The two-party vote closely corresponds to the share of seats each party wins in the U.S. House of Representatives.
Â鶹´«Ã½AV's ability to predict the outcome using the generic ballot this far out from the election using registered voters is unclear given the limited historical data from years prior to the midterm elections.
However, past data does suggest that Democrats typically need a large lead on the generic ballot among all registered voters in midterm elections to maintain a lead once turnout is taken into account.
For example, the Democratic Party averaged an 11 percentage point lead among registered voters in all of Â鶹´«Ã½AV's generic ballot polling leading up to the 2006 midterm elections, compared with an eight-point Democratic advantage in the actual 2006 House vote. In 2001-2002, the average Democratic lead among registered voters in the generic ballot was three points in an election the Republicans eventually won by four points. And in 1998, Democrats enjoyed an average six-point registered voter lead while the eventual House vote was about evenly split between Republican and Democratic candidates.
The six-point Democratic advantage among all registered voters in the current poll suggests the 2010 election could be quite close if it were held today given low turnout in midterm elections and the usual Republican advantages in turnout. (Â鶹´«Ã½AV usually takes turnout into account late in the campaign by applying a likely voter model to the data, usually beginning in September or October in the year of the midterm elections.)
Thus, at this early stage, 2010 does not look like it is shaping up to be as strong a Democratic year as 2006 was, and that could make it difficult for the party to hold onto the gains it made in the 2006 midterm and 2008 presidential elections.
Party Loyalty High
The poll data suggest party loyalty in the 2010 congressional vote is likely to be high. Currently, Republicans say they would vote for the Republican candidate by 93% to 3%, while Democrats would vote for the Democratic candidate by an equally lopsided margin, 94% to 4%. Independents are currently evenly divided in their vote preference. Democrats maintain a lead on the overall ballot given the fact that more registered voters identify as Democrats than independents.
Party loyalty is typically quite high on the generic ballot, so, in large part, the outcome is determined by the voting preferences of independents and voter turnout by party. In 2006, Democrats had a compared to the Democrats. Thus, at this early stage, it appears as though Democrats may not be able to count on the same level of support among independents they enjoyed in the 2006 midterm elections, so offsetting any usual Republican advantage in turnout could be key to holding on to the gains made in the 2006 and 2008 elections.
Bottom Line
Â鶹´«Ã½AV's initial read on 2010 midterm election voting preferences shows Democrats with a slight lead among registered voters, but one that is hardly large enough to suggest they are in a solid position with more than a year left before votes are cast.
Historically, the party whose president occupies the White House loses seats in a midterm election. Thus, the Democrats would probably expect to lose some of the seats they won in the House in 2006 and 2008. However, parties have been able to minimize their party's losses if not make gains in midterm elections when the president was popular, . Although President Obama has been quite popular for much of his presidency, his approval ratings have declined in recent weeks.
Survey Methods
Results are based on telephone interviews with 916 registered voters, aged 18 and older, conducted July 10-12, 2009. For results based on the total sample of national adults, one can say with 95% confidence that the maximum margin of sampling error is ±4 percentage points.
Interviews are conducted with respondents on land-line telephones (for respondents with a land-line telephone) and cellular phones (for respondents who are cell-phone only).
In addition to sampling error, question wording and practical difficulties in conducting surveys can introduce error or bias into the findings of public opinion polls.
