The Electoral College will determine the next president of the United States. Not polls, and not even primaries. Campaigns have to answer the question "How does a candidate get to 270 electoral votes?" Any republican candidate coming into a general election is already behind the democrat based on historically red to blue states. A fun website to see this dynamic is 270towin.com where the electoral map is interactive. Primary results can, however, help us see how voter will choose in swing states. In the end, the fate of any nominee will come down to about 10 states.
Add up the electoral votes from red and blue states, you'll see that any republican will likely receive 191 votes. Any democrat will, likely, get 217. That's right off the bat. Ten other states are "up-for-grabs", and they are NV, CO, IA, WI, OH, VA, NH, FL, PA and NC. The historical voting record of these 10 states shows that only 8 states that are true tossups. Some experts say PA is a swing state, but it's gone blue since 1992. Some say that NC is a swing state, but it's only gone blue once in the last nine presidential cycles. Only the excitement from Obama's 2008 campaign carried it into Democratic country. If we nestle NC and PA into the appropriate colors, the new count is 237 Dems and 206 GOP. This may look like a close race between the parties, but it isn't. If the republican candidate won six of the remaining 8 states, the candidate would still lose the White House. The final count would be Dems 270 to 268 Rep. If the Dem won the exact same 6 states, the final results would be 299 to 239 in favor of the Dems. Those six states are NV, CO, IA, WI, OH and VA. Based on the eight states I've outlined, or even the 10 states "experts" outline, it is nearly impossible for a republican to win in 2016.
But, I could be wrong. I'm really only basing my prediction on historical elections since 1960 and this year is breaking all the rules. It may actual boil down to the candidates. As caucuses and primaries roll out across the country, we can look at actual results of these contests to see who may likely win swing states. As of right now, two swing states have cast votes: NH and IA. In fact, many of the primaries are absolutely meaningless to general elections. South Carolina gets tons of coverage for the Democratic primary. But everyone knows SC isn't in play. The last time SC voted Dem was 1976, and that was for a native son (or at least, a neighboring native son) in an election against Gerald Ford. In fact, none of the southern states are really in play. The only primaries anyone should care about as predictors for who will win in the general election are swing states.
We can't analysis the IA caucus because the Iowa Democratic Party doesn't, or at least, hasn't released raw numbers for voter turnout. We can infer a few things though. If you look at the general election trends within IA over the last two cycles, you'll see that Obama received six thousand less votes in 2012 than he did in 2008. Inversely, the Republican candidate received nearly 50,000 more votes in 2012 compared to 2008. Total voter turnout was basically the same for both elections. If this trend continues, IA will likely go red in 2016. But that's really just an educated guess.
New Hampshire is a different story. We have actual voter turnout numbers for both parties. The total number of voters in both the Republican and Democratic primaries were 535,094. Of that total, more republicans came out to vote than democrats, but only slightly more. Of the 535,094, 53% were red voters and 47% were blue. Based on these numbers, NH would go red giving its 4 electoral votes to the republican candidate.
This would also fit into the trend of NH over the last two election cycles. Though Obama won the state in 2008 and 2012, he won by a smaller margin in 2012 than 2008. In fact, a similar deficit shows up in NH as IA. Obama received 15,000 less votes in 2012 than the previous cycle while the republican candidate received over 13,000 more votes.
Interestingly, we see a different story when we compare the voter turnout for individual candidates in NH. Of all the votes cast in the primary, Trump received over 100,000 of them. Clinton did not break 100,000 votes, only getting 95,252. So, a head-to-head contest between Clinton and Trumps in NH shows that Trump would likely win when comparing total party votes and individual candidate votes.
If you compare Trump's votes to Sanders, you get a different result. Sanders received more than 150,000 votes in the primary. This means that he would clearly beat Trump in a head-to-head contest based on primary voters. Of course, because of the Electoral College, the democrats can lose NH in the general election and still easily win the White House. Inversely, NH is a critical state for the republican candidate.
As more swing state voters cast their ballots, we can use actual turnout and votes to predict the next president. Everything else is just noise.