So I was wrong. We listened to what everyone else was seeing in the field and adjusted our Kentucky Republican primary turnout model, which we should not have done.
Our original multi-mode survey results were 0.4% from where Governor Matt Bevin ended up with on primary election day.
I think it was Mark Twain who said, “Whenever you find yourself on the side of the majority, it is time to pause and reflect.” Bevin was certainly on the side of the majority, but not the well-above-60% majority that we were hearing from “traditional” pollsters and pundits – including Bevin’s own campaign (which we had a hard time believing based on his high negatives). So we adjusted the scoring on our survey to not be 12+ points out of lockstep with what our peers were showing. It moved the Bevin ballot four points and the other races negligibly.
This same thing happened in Ohio in 2018 when we got raw results back that DeWine was ahead three points. We refielded the survey at our expense and ended up at DeWine – Cordray tied.
Never again.
Multi-mode surveys (Cygnal’s specialty) – combining live or IVR phones with SMS, email, and online panels – work, and they work better than live-phone-only polling. The Republican Primary in Kentucky on May 21, 2019, is proof of that.
Even though we have done 330+ multi-mode surveys and they were 1% more accurate than our live phone surveys, I still had doubts. Not anymore.
It makes complete sense – voters under the age of 50 don’t answer their cell phones and don’t own landlines. Those who do answer their phones in this demo don’t look like the rest of the voters in the same group. So how can we reach them? Not by calling more cell phones. Instead, we’re getting representative sample by reaching people where they are – not where they used to be.
So why aren’t more polling firms taking the multi-mode approach? It’s really time-consuming, challenging, and expensive if you don’t field your own surveys. Traditional pollsters write a script and hand it off to call centers who handle all the data processing, sample collection, and (usually) the reporting. Call centers aren’t set up to collect responses from four channels while keeping the balance across dozens or even hundreds of quota groups.
That’s what makes Cygnal different. We control 100% of the fielding process and have proprietary software, custom-built to enable near-perfect sample collection in-house. For this reason our Kentucky GOP primary survey was incredibly precise despite everyone else getting it wrong.
So I apologize for not having 100% faith in our ground-breaking polling methodology and showing you Bevin at 56% when we should have released our original Bevin-at-52% survey.
It will NOT happen again.
Here’s our Cygnal promise to you – crazy-accurate results using multi-mode channels and time-tested principles at a lower price than traditional phone research. (Yeah, that’s a mouth-full.)
If you want to learn more about the new, more-reliable approach to polling, give me a call (202) 505-5314 x700 or shoot me an email [email protected].