“When you look at their survey, they only surveyed 82 people,” Kyle Loveless with WPA Intelligence said.
He says he believes there are many reasons to be skeptical of those numbers, which were taken as a subset of nearly 500 polled statewide.
He adds that the survey doesn’t appear to verify whether those in the subset are all within city limits.
“They used a shotgun approach when they need to be using a scalpel,” Loveless said. “They didn’t take likely voters. They didn’t take registered voters.”
He also questions the methodology used.
“They used a technology called IVR, which can only be used on landlines,” Loveless said. “Well, in Oklahoma, that’s only 20 percent of the people.”
The smaller sample size also pushes the margin of error to plus or minus 10 percent for an election date with historically low turnout.