Friday, January 11, 2019

Rant: there's no secret sauce hidden in big data

Much of the media, including Sue Halpern at the New Yorker, think the revelation that GOP operative Paul Manafort shared Trump campaign polling data with a Russian agent is the long sought smoking gun -- because, well, having targeting data gave the Russians' surreptitious online interventions some magical power. Halpern reports:

Not long after the election, in an interview with “Frontline,” [proud Trump pollster Tony] Fabrizio offered a glimpse of how this data was gathered and how crucial it was to Trump’s victory. “One of the groups that we created early on in the campaign from the polling was what I called Trump targets,” Fabrizio said. “These were voters who wanted to change direction, wanted a new direction, weren’t voting for Trump, weren’t hardcore Democrats, weren’t hardcore liberals, weren’t hardcore Hillary supporters.” Knowing whom to target, where they were, and which issues resonated with them gave Trump’s digital team crucial information for its advertisements and social-media messaging. “We would report out to the senior team what markets those voters were concentrated in,” Fabrizio told “Frontline.”

Come on, that's just classic polling vendor flimflam, claiming that their piece of the work was THE essential element that won the victory. It would have been professional malpractice if Fabrizio's firm, Fabrizio, Lee & Associates, had failed to provide targeting information to the Trump organization; that's what pollsters are hired for. Yet it wasn't rocket science to come up with a profile of a plausible, perhaps persuadable, Trump voter. It's far more likely that the pollsters confirmed the intuitions of the campaign than that they came up with the secret sauce that won the election for their candidate (and the Russians' candidate).

I've worked in a lot of elections. I've worked with big data and done some targeting. I understand that elections are won with good targeting, working to influence the appropriate voters whether by persuasion or by increasing turnout among supporters who might not vote. It's always important not to encourage or sometimes to discourage the other sides' voters. But deep marketing knowledge, which is what social media can provide, is no more valuable, and perhaps less valuable, in that project than more obvious data: location, population demographics, and voting histories.

At the Washington Post, Phillip Bump has analyzed the known Russian activity on Facebook from 2015 through 2017, charted those interventions, and concluded there was no magical formula.

After all, the common understanding is that Russia’s interference efforts included sophisticated targeting of specific voting groups on Facebook, which could have made the difference in states that Trump narrowly won on his way to an electoral-vote victory.

That understanding about Russia’s sophisticated targeting, though, is not supported by the evidence — if it’s not flat-out wrong.

He shows that Russian disinformation largely went into interventions that were nationwide rather than directed at battleground states -- and that much of what has been uncovered occurred after November 2017. Bump's reporting corrects attractive myths.
We have plenty enough confirmed facts to show that Russia sought to elect Donald Trump and defeat Hillary Clinton. We also have plenty to prove that lots of people around Trump and probably Trump himself knew Russia was at work on their behalf. Why else have so many in the Trump camp lied about their contacts?

The lawyers tell us that we can't say the Trump campaign committed "treason" for having conspired with a hostile foreign power. Okay, it wasn't "treason," because the Founding Fathers knew that those in power would abuse any wide definition of that crime. But we're not wrong to see in Trump and Russia playing footsie a monstrous betrayal of the country.

But what all this does not prove or even suggest is that Russia's use of Manafort's polling data put Trump over the top. Not even hacking Democratic Party polling models did that. Trump and his Russian buddies worked to inflame and inflate our differences, to raise up racial hate and religious bigotry, to nurture cynicism and despair. But that's on us. There's no magic secret technological alternative to the long hard work of healing the violence within our society.

No comments:

Related Posts with Thumbnails