As little as a fortnight ago, the chances of Donald Trump winning the presidency were hovering around about 15%, across most of the various polls and news networks. As with any election race, the polls always appear to tighten as you get closer to the deadline and that can easily be explained with nothing more than simple mathematics.
Suppose that in the run up to election day, ten per cent of prospective voters on both sides decided to switch. Obviously, the side that started with more voters numerically, would also lose more voters numerically; that switch would increase a smaller base to a greater effect than a larger one. Likewise, if there was a pool of undecided voters and they happened to split evenly, then that would also have a greater effect upon a smaller base than a larger one.
The other explanation that is often given, as stated in previous posts, is the "Bradley Effect" where people don’t want to admit that they happen to support an unpopular candidate. I suspect that a great deal of this went on in the 1948 Presidential Election when Harry S Truman won the presidency over Thomas Dewey despite never despite never leading any of the polls conducted; including exit polls on election day. If Hillary Clinton supposedly has a difficult job in winning what amounts to a third Democratic term, then Truman seemingly achieved the impossible in winning the fifth consecutive presidential win for the Democrats.
This being said, some races like California or Texas have been firmly in their own column for a very very long time and no tightening of the polls is going to have any effect on the result, unless there is a sudden and whelming flood of votes that change sides all at once.
Mostly I’ve been assuming that Clinton should have had this election in a virtual walkover but I’m increasingly having to rethink this as polls begin to shift wildly in the closing days. It could very well be that Trump might end up winning the Presidency, despite losing the popular vote and that would be as weird a result as 1948.
America is a vast and unwieldy country. It is so big that the contiguous 48 states cover four time zones and if you include all territory, the United States uses nine standard time zones. This has the effects that as polls close at night in the eastern states, people in Hawaii are still getting back from a late lunch. What this means is that watching results unfold in America can be like a slow moving story that takes far too long and you just want it to be over.
If you can’t be bothered to sit through four or five hours where not much happens, then I have the solution which I believe is the only race that you need bother with; that is New Hampshire.
New Hampshire is not a bellwether state; nor is it a state which is firmly in one column or the other. In the 2004 election when George W Bush was reelected, if flipped from Red to Blue and has remained there ever since. It flipped Republican for the 2000 election which installed Bush and in 1992 and 1996 it flipped to the Democrats in the two elections which Bill Clinton won; before that it had been solidly Republican as far back as Nixon in 1968.
New Hampshire then, isn’t a bellwether but acts more as a leading indicator and because it happens to be one of those states jammed all the way up there in the top right of the country, the results are known pretty early on.
Even though, Clinton leads by as many as 9 points in some polls. Trump is actually ahead in others. All the way back in July, Trump was leading the polls but that switched over after the party conventions had taken place. What makes New Hampshire particularly interesting, is that even though it is the 9th least populous of the states and only gets 4 Electoral College votes, those votes are notoriously difficult to switch but if you can get those to switch over, then the chances are that votes have switched in the country overall.
In the current 114th Congress which expires next January, New Hampshire sent one Republican and one Democrat to both the House of Representatives and the Senate; so if that’s any indication then the race for the presidency in the state could equally go either way.
The thing to remember about the race for the White House is that it is not one election but 51 elections all going on at the same time (DC isn’t a state but for the purposes of a presidential election it is treated as though it were one). About a third of all of the election races are almost foregone conclusions so you don’t need to worry about those; some of them like Florida dance on a knife edge and are impossible to predict; so it is places like New Hampshire which actually give the best indication of what the overall picture might look like.