Yesterday, Nate Silver put out his latest model showing Trump with an official lead in the electoral college predictions in this tweet. According to Nate, it's basically over, Trump has already won. Understandably, this scares many people, and at the same time makes others very happy.
I am neither of those things. What am I? Glad you asked.
I'm skeptical, patient, and calm. Everybody, stay calm.
“BUT HOW CAN YOU BE CALM! TRUMP IS GOING TO WIN! THAT'S __________ (insert your preferred descriptor here: Great, Terrifying, Amazing, Awful – mad-lib political analysis, truly cutting edge) AND THAT IS A REALLY BIG DEAL! IT'S YUUGE!”
Ok, I hear you. It's just that... well... it doesn't actually matter. At this stage in the game the national polls mean almost nothing. It will be late September before we have a better snapshot of the likely outcome of this election and even then no model can fully predict reality.
WHAT AM I TALKING ABOUT?
What Nate Silver at www.fivethirtyeight.com posted was some good old-fashioned “click bait.” Post a headline that excites people, either in a good way or a bad way, and people are likely to click on it, share it, repost it, (and then they tell two friends...) and suddenly you've earned one million clicks today. Your site sponsors are happy and you justify your role as the interesting prediction guru guy.
Good for Nate, he's done very well in that role. But, again, what Nate Silver says on July 26th does not an election outcome make.
Let's use some context here, all thanks to www.realclearpolitics.com and their endless fun for those of us that think fun is spotting trends and relationships between historical elections and current elections. It's fun. Trust me.
On September 6, 2008, John McCain was polling at 44.2%, and Obama was at 46.6% in the national polls. On September 7, 2008, McCain was up 46.7% to 45.7%. McCain would get as high as 48%, and a full 3 point lead over Obama during mid September. On election day, Obama won with 365 electoral votes. Of the last four elections this one in particular evidences my point: July polls are basically useless.
Through a large part of the 2012 election the news media and polling represented that Obama and Romney were basically tied. Every news outlet was talking about 2012 being our first real tied presidential race, in modern time. Then, election day came. Obama won with 332 electoral votes. It was closer than in 2008, but never in question.
In 2004 Kerry spent most of the summer with a slight edge over Bush in the polling data. On election day Bush won with 286 electoral votes, despite the poor summer polling numbers. That was a close election, but Bush still won and fairly easily.
In 2000 the race was tight from the word go. Gore was never that popular but was the clear front runner for the Democrats and Bush was the only GOP candidate to ever poll over 30% in the primary. (Interestingly Elizabeth Dole was the only other GOP candidate to poll over 20% in that cycle.) In July of 2000 Bush was up on Gore 50% to 39% in the polling. It tightened in the fall, but Bush always remained slightly ahead overall.
Of the last four Presidential elections, the 2000 election is the only one where the polling in July was indicative of the final outcome in a meaningful way.
Get it? The polls don't mean as much today as we give them credit for. They can be interesting, they can be used in fundraising, and to whip up the base, but ultimately they are only a snapshot in time. Nate Silver has fantastic models – the best models, you can't believe these models- sorry, but really they are good. They are not perfect. More importantly, they are going to change.
The bigger problem is that a national poll doesn't spell out local feelings and doesn't lay out electoral college standings. Just ask Gore from 2000 which number is more important, the national popular vote or the electoral college? He'd tell you the same thing I will: It's the electoral college.
Add to the lack of nuance the complication created by a viable third party candidate and you've got nothing but garbage to look at with national polls. One thing we know for sure right now is that the polls will change. The pollercoaster is never a smooth ride, and 2016 is no exception.