New company promises more accurate polling results

By Jeff Beimfohr

Click here for updates on this story

    SHREVEPORT, Louisiana (KTBS) — It’s election season. And we’re inundated with political polls.

But as we head down the stretch to election day a new method of statistical sampling has originated in Shreveport that could change the face of political polling forever.

Vantage Data House is the brainchild of two men — Bishop George and Jim Kitchens. Both worked closely with Shreveport’s BRF Entrepreneurial Accelerator Program to get their business off the ground.

“Some of the algorithms they use, the technology behind the whole platform, it’s really disruptive in the polling industry,” said Julie Gilley Milam of the BRF/EAP. “And we’ve seen they’ve been very successful in the polls they’ve run. It’s intriguing to see all that data in one platform.”

During Louisiana’s 2023 elections, Vantage Data House accurately predicted 97% of races up to five months out from election day.

“We adopted a polling methodology I was familiar with, that had never been used in politics,” said Kitchens. “Pew Research used it in international polling. The Census Bureau uses it, and we do very large samples in a state, and it allows us to drill down to the congressional level, the legislative district level, even the county level.”

The Vantage Data House method is called “Authenticated Predictive Projection.”

“It’s a mathematical algorithm that says we look at the makeup of a district on five factors: where you live, what’s your age, what’s your party registration, what’s your race and what’s your gender,” Kitchens said.

From there, Vantage Data House can drill down on a district. It’s been field tested in five states and academically reviewed by the Universities of Kansas and Missouri.

“We did our first big test in Louisiana, where we had 30 races we looked at. In some we would actually do a traditional poll to make sure the numbers matched up and our projections were working,” Kitchens said.

And they were. Businesswise, Vantage Data House switched up the norm.

“We simply changed the business model,” said George. “Instead of individuals buying a $15,000 to $80,000 poll, we switched it to a membership model where you are just leasing the information from us.”

“Instead of paying all that per poll, one snapshot in time,” explained George, “you become a member with us for a fraction of the cost and you get access to 130,000 polls for just $416 per month.”

Vantage Data had to automate to accommodate up to 40,000 respondents per state.

“The margin for error is so small because you have asked so many people their preferences. It’s just that much harder to get it wrong,” said attorney and political analyst Royal Alexander.

And all that information is update every single day. Could Alexander see it replacing traditional polling methods?

“I could, because it’s so much cheaper, “said Alexander. “We’re no longer sample size 400. We’re now sample size 4,000 or more. Dr. Kitchens has been doing this polling and this type of statistical sampling for 50 years, so his methodology is as solid as there is.”

By way of example, KTBS asked Vantage Data to look at the congressional race in Louisiana’s new District 6.

On the day of these interviews, George explained: “We go to District 6 and see right now Cleo Fields is winning with 42.9%, Guillory is at 24.8%. We created the first-ever automated polling experience ever invented.”

“And it doesn’t even have to be a political poll, it could be a poll about any type of information,” Milam said.

“But nobody’s ever approached politics this way,” said Kitchens.

Milam agree. Vantage Data House could change the face of polling and emphasized the reasons behind their interest.

“We’re looking for something that is going to be high growth; something that is going to create jobs in North Louisiana. But we’re also looking for something that’s innovative, maybe disrupting the industry; something that is unique.”

Please note: This content carries a strict local market embargo. If you share the same market as the contributor of this article, you may not use it on any platform.