PRINCETON, NJ -- One year from today, U.S. voters will head to the polls to elect all 435 members of the U.S. House of Representatives. Gallup measures voting intentions in midterm elections using the generic congressional ballot, which has proven an accurate predictor of the eventual vote in midterm elections.
Gallup's most recent test of the 2010 elections, from an Oct. 1-4 poll, showed 46% of registered voters saying they would support the Democratic candidate in their local district if the election were held today, compared to 44% who would vote for the Republican candidate. That was a slightly better showing for Republicans than Gallup's prior test in July, which had Democrats up by 50% to 44%.
In terms of analyzing the generic ballot for clues as to which party has the advantage well before an election, the size of the lead among registered voters offers arguably more insight than the simple fact of which party is ahead. Given the usual Democratic advantages in party identification among the general public, it is rare for Republicans to lead on the generic ballot among registered voters. This was the case even when Republicans were the majority congressional party from the mid-1990s to the mid-2000s.
But the reason the Republican Party is competitive in congressional elections is that Republicans generally turn out to vote at higher rates than Democrats. Turnout is crucial in midterm elections, and with at least 80% of Americans registered to vote but only about half that number likely to vote in the midterm elections, registered voter and actual voter preferences can differ by a lot.
As a starting point, if Democrats have close to a double-digit lead among registered voters, they are virtually ensured of also having a lead among actual voters -- whatever turnout happens to be on Election Day.
From that point, things become a little less clear-cut. Smaller Democratic leads do not necessarily mean the party is losing, but rather that the eventual election outcome will be more dependent on turnout. In general, the closer the registered-voter results get to an even split, the better Republicans can expect to do, given usual turnout patterns.
Gallup analyzes turnout patterns in a given election by applying its "likely voter" model to midterm results. However, since it is unclear how reliable likely voter models are in identifying likely voters in a low-turnout election long before Election Day, Gallup usually does not begin to estimate likely voter preferences until September or October of a midterm election year.
But early generic-ballot results based on the sample of registered voters do still give reliable -- if not highly precise -- indications of how the election might go.
For example, through much of 1974, 1982, and 2006 -- highly successful years for the Democratic Party -- the generic ballot showed consistent, huge Democratic advantages among registered voters.
In 1994, leading up to a Republican landslide, the generic ballot suggested it could be a very promising year for Republicans, as the party actually enjoyed a slight lead among registered voters in several polls. When the likely voter models were applied in the final months of the campaign, Republicans held solid leads among this group of voters.
In years such as 1998 and 2002, the early reads on the generic ballot suggested the elections would be competitive, and the eventual outcome would largely depend on turnout. Those two elections were unique from the standpoint that the president's party (Democratic in 1998 and Republican in 2002) gained U.S. House seats in the midterm elections, bucking the historical trend. Turnout proved to be pivotal in 2002 as the Democrats' five-point lead among registered voters turned into a six-point deficit once likely voter preferences were measured (the actual vote on Election Day showed a five-point Republican advantage).
Translating Votes to Seats
The aggregated national vote in and of itself does not directly determine the election outcome since that vote is spread out over 435 districts, some of which are highly competitive while others are not competitive at all. The focus of House election results is the tally of the number of seats won by each party. Still, the national vote has proven to be highly correlated with the eventual seat distribution by party in the House of Representatives. This allows Gallup to use the generic ballot to make predictions about the eventual Democratic-Republican seat distribution and the likelihood of either party's having control after the election.
Gallup has developed a statistical model to attempt to estimate the House seat distribution based on the national vote. This model takes into account historical voting patterns in midterm elections from 1946 to 2006 and structural factors such as the president's party and the majority party in Congress at the time of the election. In 2006, Democrats enjoyed a seven-point lead among likely voters in the final pre-election voting estimate, suggesting a range of seat gains (given the model's prediction error) of 11 to 58 seats, with a 35-seat pickup the most likely outcome.
The Democrats needed only 15 seats to regain majority control from the Republicans that year, which seemed almost assured given the poll estimates, and the Democrats ended up gaining 31 seats in that election.
Given the Democrats' current majority status (they now hold 257 seats), it is unlikely that Republicans in the next election will win the number of seats necessary for them to become the majority party. Although Gallup will not be able to predict each party's share of the congressional vote with a great deal of precision until it applies likely voter models next fall, the statistical model used to predict seat distribution by party can use historical data to present a number of scenarios in terms of seat distribution, based on the ultimate vote by party.
Gallup's model shows that Democrats could lose the overall 2010 congressional vote and still retain majority-party status. With as little as 51% of the two-party vote, Democrats could expect to hold almost all of their current seats (the model would predict 251 Democratic seats held after the election, with a range of 240 to 262). And with as little as 48% of the vote, Democrats could still reasonably be expected to retain the majority of 218 seats (although they could conceivably still lose the majority at this level, given the model's error range).
Bottom Line
Throughout the 2010 campaign, Gallup will measure voter preferences for the U.S. House, using the generic congressional ballot. A review of past midterm election-year data suggests that even early predictions can give insight into how next year's elections might go. The size of the Democratic lead among registered voters in the months ahead will be a key indicator of whether the election will be competitive. If Republicans are close to being tied or are ahead, it will suggest the possibility of a strong Republican year unless Democrats have an unusually strong and disproportionate turnout advantage.
In Gallup's most recent update, the Republicans are within two points of the Democrats among registered voters. The closeness of these results suggests that voter turnout, which Gallup will begin to assess next fall with its likely voter model, will be an important factor in the ultimate 2010 election outcome. Once probable voter turnout is known, Gallup can provide greater insight into possibilities for changes in party division in the next House of Representatives.
The possibility of a strong Republican year would be consistent with history. The president's party usually loses seats in a midterm election year. The popularity of the president is a factor in the size of those losses. In a forthcoming analysis here on Gallup.com, Gallup will look at some important indicators beyond the generic ballot -- including presidential job approval -- that can give additional clues as to how the midterm elections will play out.
