Poll consumers should not use nationwide polls to guess what is likely to happen in U.S. Senate races this year. Likewise, no matter what the outcome, these Senate races will not predict what will happen in the 2016 Presidential race.
Senators are elected for six year terms. Elections for those seats are divided into three classes, and the Senators elected by each state are in two of the three classes. As a result, approximately one-third of Senate seats are contested each two years in general elections. Seats contested this year are part of Class II. Due to vacancies since the last general election, two states (Oklahoma and South Carolina) have two Senate races this year.
Class II states are not representative of the country as a whole. On most grounds, they are the least representative of the three classes.
For example, Class II is much smaller than the other two "classes" of Senate races. While states in the other two classes each represent over 70 percent of the votes cast for President in 2012, Class II states represent only 52.3 percent. No Senate elections are being contested this year in five of the seven most populated states: California, Florida, New York, Pennsylvania and Ohio.
Second, Class II states were much more supportive of Mitt Romney than the nation or states in the other two classes. Nationally, Barack Obama defeated Mitt Romney 51.1 percent to 47.2 percent. But Romney won a majority in the Class II states, while Obama won majorities in the states assigned to the other two classes.
It is also worth remembering that the Senate seats up this year were last contested in a regular general election in 2008, the year of Obama's initial election. Several Democrats up for election this year were elected in what had been Republican or Republican-leaning states in previous Presidential elections. In fact, of the Senate races that have been held in Class II states in the two general elections since 2008, 20 of 37 have been won by Republican candidates.
Post-election analysts should be similarly cautious about drawing major conclusions about what the 2014 Senate races portend for the 2016 Presidential election. Many assumed Democrats were well-positioned for Presidential elections in 1988 and 2004 based on the outcome of Senate elections two years before. Republicans were equally gleeful of their chances going into 2012 because of their success in the 2010 Congressional elections.
Plenty deserves to be said about the 2014 Senate elections. They will, no doubt, play a major role in the legislative record of the next two years. But it would be a serious mistake to suggest that these 36 elections represent the nation as a whole or predict the likely outcome of the 2016 Presidential contest.