With less than a day to Mid-Term election day polls are narrowing and pundits are coming up with all sorts of reasons why the Democrats will hold the House of Representatives. Most of the pundits – who are pretty universally Democrat-leaning – have conceded that the Republicans will hold the Senate but the House seems to offer the best chance for a Democratic victory. I wrote a month ago that I was not convinced that the House would swing Democrat and I ascribed that to the popularity of President Trump.
After a month of hectic campaigning, I see no reason to revise that thinking. In fact, if anything, the Republicans seem to have been galvanized by the Trump rallies, the Kavanaugh outrage and the poor people in the caravan(s) heading for America’s southern borders.
Early voting is up. Much higher than it usually is for mid-term elections. And in states where people are registered Democrats or Republicans, it appears more registered Republicans are voting early than Democrats.
2016 demonstrated that the polling model – generally and specifically – is badly broken. However, the question is whether the breakage is valanced in any way. The reason why polling is broken is that people are not as willing to answer their phones as they were twenty or even ten years ago and large numbers of people no longer have landlines. Their mobile is their only phone. And you cannot legally autodial cell phones. Ten years ago you could argue that the cell revolution skewed young and therefore young people were likely undersampled. In theory, young people were more likely to support Democrats thus the cell issue was valanced with Democratic Party support underreported.
To compensate for this, polling organizations overweight their samples to try and capture the missing Democratic support. It is a dark art and one which grows increasingly unreliable as more and more people become unreachable or unresponsive to pollsters.
The other huge change in the last decade is the waning of the influence of mass (and hugely liberal) media. Newspapers are in their death throes, network television is fighting for audience with cable and both are being sidelined by everything from Netflix to You-Tube to Facebook. Where people get their news and where they see advertising has profoundly changed.
A decade ago, in a tight race, a Party might make a strategic TV ad buy to haul its candidate across the finish line. Parties are still doing this but it is not at all obvious that races can be swung with a million dollar last-minute ad buy.
So how will the 50 or so House races which matter be decided?
I am pretty certain that many of these races will come down to which party gets out its vote, the good old-fashioned “ground game” with some information age bells and whistles.
Which is the reason I think the best indicator of tomorrow’s result is the surge in early registered Republican voting. The ground game begins long before the election and one of the key strategies is to get your identified voters to the polls as early as possible. The logic being that that reduces the amount of work which has to be done on election day. The Republicans spent a lot of time and a lot of money identifying their supporters in battleground states in the 2016 election. Those lists are fresh and available to candidates and campaigns.
E-Day tactics have not changed much in fifty years. A successful campaign will target its supporters who have not voted and get them to the polls. Phone calls and, even better, door knocks can make a huge difference in the final result. It is hard, not very glamorous, work.
But the Trump machine has done something no mid-term campaign in history has done: it has managed to get its activist core to register themselves as activists. This was the brilliance of the Trump rallies.
I very much doubt that a single mind was changed as Trump rocketed from hanger to hanger with Air Force One doing double duty as the world’s most exclusive stage prop. The rallies were all about affirmation, being part of something, knowing you were not alone. But they were also a brilliant way to collect GOTV data.
To go to a rally you had to register, online, for a ticket. Literally, tens of thousands of hard core Trump supporters registered. Over all I suspect well over a million people willingly gave up their emails and phone number for a chance to get a ticket to the Trumpaloza in their state. Which meant they self-identified as Trump Republicans. And it also gave individual campaigns lists of people who were likely to a) vote early, b) be willing to work on the GOTV effort, c) be ready, willing and able to knock on the doors for the E-Day effort.
Trump’s 2020 campaign manager, Brad Parscale, is data driven. But he also understands that data without action is pointless. The rallies with their registrations have given the Republicans an army in the battlefield states the likes of which the Republicans have never had before.
We’ll see on Tuesday if it is enough to add seats in the Senate and hold the House.
It will depend on who shows up.
Update: Tuesday Afternoon. Reading anecdotal reports on turnout. The general trend seems to be high turnouts – at or better than Presidential – in red areas. Low turnout reported in some blue areas. You have to read a lot of reports to get a decent picture and there will be confirmation bias all over the place; but those trends are what will be needed for a red wave.