In an interview with the Britain's Channel 4 reported by the Guardian, Donald Trump's still relatively new campaign manager, Kellyanne Conway, offered her explanation for why polling that shows Trump trailing badly is wrong — and why she's convinced that the campaign will nonetheless win.
Her explanations did not convince the media, in that they did not convince me.
First, on the polls:
"Conway insisted that Trump’s support was not reflected in polls because of the perceived social stigma of supporting the Republican nominee. ‘Donald Trump performs consistently better in online polling where a human being is not talking to another human being about what he or she may do in the elections … it’s become socially desirable, especially if you’re a college educated person in the US, to say that you’re against Donald Trump,’ said Conway."
This is the ol' Bradley effect, the idea that people misrepresent their true positions to pollsters because of social stigma or for some other reason. The idea Conway presents is that it's hip for college-educated Americans to say that they're anti-Trump, and that when a pollster calls and asks a voter who they support, those meek college-educated Trump backers demur or pick Hillary Clinton. Which is why Trump does better in online polls, she says, where there's no actual human calling.
I looked at this last week, when Conway first took over as campaign manager. I grabbed online and live-caller polls from the beginning of the conventions and looked at how the results compared. Trump's average support in both types of poll fell after the conventions — but online polls also had Trump doing worse than live-caller ones.
Our resident polling expert Scott Clement looked at several recent online and live-caller polls, too, and found that the average margin between Clinton and Trump was only 1 percentage point different between the two types of polls (44 to 40 in live phone polls; 44 to 39 in online).
Regardless, Conway intimated that the campaign had a way to reach those timid voters or people who would back Trump but might not go to the polls.
"Conway insisted: ‘We give people a comfortable way to express that maybe they don’t want to vote this year and why that is.’ She described her method as ‘proprietary’. She said that as a result, she could reach these undercover voters ‘in many different ways’. She said: ‘We go to them where they live, literally.’"
Campaigns do lots of things that they don't tell the media about, for obvious reasons. Campaigns also often have ways of campaigning that are unique to the campaign team, though those unique aspects are mostly at the edges. The broad strokes of all campaigns are generally the same, and when you "go to [voters] where they live, literally," that means literally going to where they live and knocking on their doors, as part of a field effort.
But Trump's campaign isn't spending money on field campaigning. It should be staffing up to contact voters before early voting begins and to have volunteer bases in swing states before the final weekend. It's letting the Republican Party do that.
So what's this proprietary system? In July, Trump spent $8.4 million — nearly half of the total amount it spent — on its digital effort, which suggests that the campaign plans to do voter turnout and contact online, versus on the ground. That fits with Trump's strategy in general: lots of tweets, not many roundtables.
To a layperson, this seems to be a sensible idea. It's not as sensible as it seems, particularly given Conway's stated goal.
Campaigns should spend time and money doing one of two things: convincing voters who will definitely vote to go vote for their candidate and convincing people who will definitely support their candidate to go vote. That's it. A broad strategy of advertising online can be good for the first effort. You can target people on Facebook with a specificity that is hard to replicate using a TV spot. Trump is working with a vaunted digital-targeting firm that can make that process of targeting easier. (It partnered with Sen. Ted Cruz (R-Tex.) in the primary which ... wasn't a complete success.) Facebook allows campaigns to match supporters to voter lists, ensuring that particular individuals can be targeted. (The more specific the targeting, incidentally, the higher the cost.)
That targeted advertising is very different from getting people to vote. You can go to where a voter lives by way of their Facebook accounts and show them an ad you want them to see — but you can't get them to vote. You can put all the lawn signs you want on a particular voter's lawn, but how do you get them to actually cast a ballot? Because voting is a habit and less-frequent voters are less likely to vote (by definition), you may be showing an ad to someone who isn't going to bother voting. Someone at the door could convince them (as research proves)! A Trump ad on Reddit probably won't.
These are smart people at Trump Tower, and this could be Conway presenting an un-rebuttable case for the campaign's path to victory recognizing that the odds are stacked against them. Or it could be the path to victory they actually see, somehow. Most of Trump's senior team hasn't run a presidential campaign before, which may mean that the latter is more likely.
Campaigns, especially campaigns with money, are constantly fending off out-of-the-box paths to campaign victory. The best indicator of how rarely breaking all the rules of campaigning works is by asking winning candidates how they won. In nearly every case, they did it the old-fashioned way. And with that in mind, campaign managers that move on to the next campaign tend to adhere to another core tenet of campaigning: "If it ain't broke."
This post originally said Conway's interview had been with the BBC. A bit of American stupidity there on my part; the post has been updated.