For five weeks, RAND has surveyed members of the American Life Panel (ALP) leading up to today's midterm elections. These surveys asked ALP participants about their voting intentions, as well as their beliefs and opinions about a number of important current policy issues.
What sets the RAND surveys apart from other public opinion surveys is that we surveyed the same people each week. This allowed us to observe changes in voting intentions over time as we drew closer to election time. Furthermore, many of these ALP members had participated in the panel during the 2012 presidential election. This allowed us to link their responses now to responses then and observe changes over a longer time period.
For the final weekly survey ahead of the 2014 midterms, we surveyed 2,872 members of the ALP between 12:01 a.m. Oct. 26 and 11:59 p.m. Nov. 1. As with other surveys, we weighed responses to ensure that results are representative of the population, matching to the 2014 Current Population Survey. Margin of error is calculated as 1.96 times the standard error. Additional information about the methodology is available in our methodology report.
Our Latest Predictions
Although the Republicans have a strong chance of maintaining control of the House and possibly even gaining control of the Senate, our survey results suggest that while individual races may vary, the support for Republican candidates nationwide may be less than the overall support for Democratic candidates.
Our results suggest that 50.5 percent of voters (margin of error +/- 2.3 percent) in the nation will vote for a Democrat and 42.4 percent (+/- 2.3 percent) will vote for a Republican in the upcoming election for the U.S. House of Representatives, and that 47.4 percent (+/- 3.3 percent) will vote for a Democrat and 44.9 percent (+/-3.3 percent) will vote for a Republican in the upcoming election for the U.S. Senate.
Since our survey is conducted on a nationally representative sample of Americans, we are able to gauge the political leaning of the country as a whole. Our results are in contrast to those from other generic congressional polls, which suggest a stronger position for the Republicans. Several possible explanations exist for the difference.
First, our methodology differs from what is normally used, because we ask about the percent chance of voting and the percent chance of voting for each party, rather than simply whether they are likely to vote and who they will vote for.
Second, response rates to our surveys by political party may differ from those for traditional political polling for a variety of reasons. For instance, it is possible that respondents in the RAND ALP are more willing to respond to political questions than in traditional cold-call political polls because of the longitudinal, ongoing surveying of these individuals on a wide range of topics, not just on political questions during an election period.
As we noted earlier, today's election will be decided district-by-district, and state-by-state. Yet the results may not be indicative of the country as a whole.
Katherine Carman is an economist and Michael Pollard is a sociologist at the nonprofit, nonpartisan RAND Corporation.
Commentary gives RAND researchers a platform to convey insights based on their professional expertise and often on their peer-reviewed research and analysis.
