Data-driven discourse has the power to break down the silos that separate us. It is the escape hatch for the echo chamber.

The day the polls were all wrong

The thesis of WIRED’s June 2016 article “The Polls Are All Wrong” seemed hyperbolic at the time. Then November 8th rolled around.

Interviewed by data.world CEO & co-founder Brett Hurt, Dan Wagner, Founder and CEO of Civis Analytics, explains that the polls were all wrong for a surprisingly simple reason: polling methods have remained static as modern technology advances, causing a dramatic decrease in the representativeness of polls.

“The industry has been aware of some of the problems, but hasn’t built a technical replacement for how you do it at scale… that’s what we’ve thought about a lot at this company; how do you create these primary replacements for survey research, for both political campaigns and businesses, that meet the standards of today without the problems of yesterday.”

It used to be that when you called 100 people on the phone, 10 would eagerly respond. Now only one or two may respond, and these one or two are extremely different than the average voter.

But if the polls really are all wrong, how do campaigns know who to target?

“The Trump campaign, interestingly, assessed that digital identity and digital community were hugely important — much more important than TV, much more important than volunteer hours on doors, much more important than phone calls. They knew that people wanted [digital community], and they were able to play into that. It was a mistake on our side not to realize that.”

Garbage in, garbage out

Every day, people are on the internet creating their own realities in terms of what types of information they choose to consume, and what sources they choose to consume it from. They are forging their own communities, which create and alter that reality on a daily basis. To be effective, campaigns now have to either be the creator of these digital realities, or penetrate the digital reality created by a more effective campaign.

“That will be the fundamental challenge for the Democrats over the next four years… that you have a set of people that exist in an isolated digital content silo, where their assumptions about people, their beliefs, and their community is completely segregated from the rest of the country.”

So how do we break down the silo?

“In order for a campaign to fight this, they first need to acknowledge it, then understand the digital topography of why it’s happening, and then build the means of breaking that silo… or else they’re screwed”

In our brave new world of alternate facts and fake news, this is no small task. Especially when social media titans like Facebook not only create these digital content silos, but also perpetuate and thrive off of the resulting segregation.

“Facebook’s business model is basically to build the echo chamber. The fundamental human condition they’re working with is loneliness — you log in to participate in a community of other people and don’t have to feel lonely for a minute.”

Because echo chambers fulfill the human needs of validation and connection, they drive higher usage of hours and intensity of use on social media platforms, thereby creating an inherent incentive for Facebook to maintain this model.

Dan believes that it could be possible to treat political content differently from other organically customized content on Facebook and other platforms as a means to prevent digital silos, but only time will tell whether that will be an effective means to counter this phenomenon. “Or,” he states, “society will just have grow up and learn how to live with this.”

Data as a source of truth

In the absence of data, humans make decisions based on what they feel, what others tells them, or what they can glean from minimal experience. But experience is only a fraction of the data you can collect through collective observation. That’s why Dan describes data as the “grease of decision-making,” and why the members of the Civis Analytics team proudly describe themselves as “harbingers of truth, masters of data science.”

“We do tons of work in the public sector, and it’s horrendous sometimes to just get basic information… If you had the data, and that data could be shared and manipulated, you could create a basic awareness of what’s going on in the public sector. You could provide opportunities for people to extract that, and provide basic metrics around it, and then ideally in 2–5yrs, provide the opportunity for data scientists to use it.”

Through the proliferation of open data and the subsequent democratization of access to a shared source of truth, we can begin to reverse the echo chamber effect that has fractured our society into disparate factions.

“The important thing, whether it’s a campaign or a company, is to have a theoretical and empirical discussion around how are things changing, where are things changing, and what do we need to build toward.”

This data-driven discourse has the power to break down the silos that separate us. It is the escape hatch for the echo chamber.

Watch the full interview with Brett Hurt and Dan Wagner below.


Collective Data Empowerment breaks down the silos between your data people and everyone else. But how do you get started?

Read how to awaken your hidden data workforce with this complete guide.