As Republicans gear up for tonight's debate, lots of people are still pretending national polls about primary elections actually mean something. Recent history tells us this simply isn't true. Past front-runners at the same stage of presidential election cycles have gone on to flop once ballots were counted:
None of those four ended up winning the nomination. Donald Trump's 16.9-point lead in the latest Real Clear Politics average is right in line with the leads those other front-runners lost.
It's worth noting here that we aren't even comparing the "same stage" of the cycles accurately: In 2012, four states held their contests before Feb. 1, which is when the 2016 Iowa caucuses will start the process. Twelve states -- nearly a quarter of them -- held their primaries or caucuses before March 1 in the previous cycle; only four will do so this time around. So, past front-runners faltered even though they were closer to being before the voters than the current field is. It stands to reason change is more likely now, if only because there's more time for it.
But more important, national primary polls aren't predictive because we don't vote in a national primary. At FiveThirtyEight.com, Harry Enten explains that Trump's big lead in the national polls means only that he could wind up with anywhere from 8 percent to 64 percent of the primary vote -- which is to say, his big lead in the national polls means exactly nothing.
Could Trump win? Yes. Could he end up with a smaller share of the vote than fourth-place Ron Paul won in 2012? Also yes.
Of course, you'll see and hear a lot of people over-hyping Trump's national poll results. They generally break down into three groups: People who support Trump; people who want to use Trump to get better ratings/clicks; people who really want a Democratic president. (Naturally, there could be some overlap among these groups.) I'm not sure any of them should really be counted on to tell us anything accurate about what's going on.
Far more meaningful is what will happen in individual states. That's a little harder to gauge, in part because we're only now beginning to get a sufficient amount of polling to give us a clue how things stand even in the early states of Iowa and New Hampshire. The fourth state in this cycle, Nevada, hasn't had a poll that's counted in the Real Clear Politics average in two months.
Even then, it's important to know that not every state is created equal when it comes to determining party nominees. The early states like their time in the sun, but they don't necessarily shape the contests that follow. Take Iowa: Since the first GOP caucuses in that state in 1976, there have been seven contested races (not counting incumbents running unopposed). Only three times has the Iowa winner gone on to win the GOP nomination (although Mitt Romney lost by just three-thousandths of a point in 2012). New Hampshire has been more predictive, with its winner taking five of seven contested races in the same time period.
At National Journal, Ron Brownstein names nine states that could be key to sorting out a contest that increasingly focuses on Trump, Ted Cruz and Marco Rubio, because of the states' combination of evangelical and college-educated voters. Those states are South Carolina, Georgia, Virginia, North Carolina, Michigan, Illinois, Florida, Ohio and Missouri. All but South Carolina vote between March 1 and 15 -- meaning they're even further away from casting votes than Iowa or New Hampshire.
We are approaching the time when candidates such as Cruz and Rubio will start taking more chances and making bigger bets with their campaign money and organizations. The nature of their campaigns, and those of others, will change. And some of those others will realize they aren't going anywhere, which will change the race as well. In short, we still haven't seen the decisions and moments that will decide the Republican nomination, and neither have the people talking to opinion pollsters.