Its a swing state, chock full of independents, so you get the pulse of the average independent, assuming they're politically relevant this time around. Its mos

Its a swing state, chock full of independents, so you get the pulse of the average independent, assuming they're politically relevant this time around. Its mostly a white state with whatever implications you have for race in the election (hillary won in '08 for racial reasons but 99.9% of black people everywhere else voted for Obama with a predictable result ... and likewise this time around Hillary isn't going to get the black turnout that Obama got from that racial block of 99.9% D voters both in the primaries and in November). Its very small and proportionately politically active so the residents tend to get involved more than a large empty state so at least in theory it predicts how politically active people across the country will tend to vote. Wikipedia claims a win results in a historical average boost of 27% of total primary votes. Every other election comes after the news media pounds who is a winner and who is a loser into the brains of votes which surely has some predictive capability looking at an un-influenced voter. Barring special racial situations or whatever (see '08) it historically has strong predictive capabilities. Note that you can "win" the primaries like Bernie did and still have your nomination stolen from you. I'm sure the Bern victims will be just as motivated to vote for Hillary as they were to vote for Bernie, I mean the Bern victims sure are a big bunch of globalist corrupt banker supporters, LOL, so they'll fit in perfectly with Hillaries supporters. Data from NH does predict excitement and likelihood of voter turnout and so forth which has certain implications in November. All 30% and then some who voted Trump are almost certain to vote in November, the 30% Bern victims well maybe not so much. Good predictive capability often has nothing to do with how much the results or methods are liked.