The Voter Turnout Guess

Predicting 2022, Part 3: Pollsters often evaluate different turnout models when conducting their research, but they are more guesswork than the product of data science. In part 3 of our series, we discuss voter turnout models and unveil the most reliable data point that we have seen to help make this guess a more educated one.

by David Burrell, CEO at Wick

3 min read

INTRODUCTION

Over the course of these three bands, pollsters will monitor and adjust their turn-out models based on a number of factors, not all of which are data-centric in nature. Relying heavily on instinct, experience and awareness of key issues, pollsters pay close attention to the vicissitudes of registered voters’ passions and enthusiasm to build and tweak their turn-out models. Unfortunately, there are no magic questions or any external data sources that can pinpoint for a pollster what the turnout will be among key demographics, so instead we rely on “models” for different election day scenarios. These models get easier to tweak as early voting starts and real data comes in, but the models are highly dependent on objective analyses of what cohorts might be turning out at higher or lower levels due to the political environment.

This is going to drive turnout.

Pundits and the political class use this phrase when they identify particular “super issues” voters deeply care about. For example, in the coming midterm election, we have to consider and factor in the Supreme Court’s Dobbs ruling, inflation, gas prices, woke policies, violent crime, ongoing migrant and border challenges, and interest rates among many other “super issues.” These and others fire up certain cohorts which turn registered voters into likely voters. As more issues surface, the turn-out models become more complicated and fluid, requiring careful balancing along the way. For example, will falling gas prices assuage certain cohorts’ concerns more or less than increased visibility of the migrant crisis as a result of Governor DeSantis sending migrants to Nantucket? Every day between now and election day is liable to produce events that tend to increase or lessen enthusiasm for one or more cohorts.

How can we best predict turn-out models?

In our experience over the past 14 years of producing polls at the federal, state, and local levels, we have identified and will now religiously apply objective, albeit anecdotal, feedback collected by thousands of individuals who knock on doors for candidates or in support of particular issues. The information these individuals gather by speaking one-to-one with registered voters is invaluable to fine-tuning our turn-out models. Door knockers converse face-to-face with registered voters and are able to gauge their enthusiasm and propensity for voting in any of the three bands noted above. 

In September 2020, before the last presidential election, one of Wick’s partners led a massive, multi-state door-knocking initiative and emphatically argued that the polls showing Biden up by 7-10 points were overly optimistic and that Trump had garnered far more support among likely voters. We factored this feedback into our turn-out models and our polls were some of the most accurate in the industry. When all the votes were counted, our partner was spot on. Just a couple of months later in Georgia’s two run-off elections for Senate in January, our partner concluded based on their door-knocking campaign that the passions of likely Republican voters had dissipated significantly from the November time frame due to concerns over voting integrity and January 6, and both incumbent Republicans failed to win their elections. Our polls were within the margin of error, but we neglected to factor in this important anecdotal input. We won’t make that mistake again.

Straight from the ground and into your hands

In the final weeks before the 2022 midterms, we are working closely with our door-knocking partner to ask voters about the critical “super issues” discussed above to determine which ones are most likely to impact our evolving turn-out models. Will the Dobbs decision, for example, inspire and fire up Democrats and other pro-choice registered voters to show up at the polls, and in what volume? Will inflation fire up Republicans and other pocket book voters in larger numbers than they otherwise would in the absence of these issues? As we apply our extensive experience and instincts to our turn-out models and scenarios, we will aggregate and balance all of these “super issues” across the board to ensure our polls are as accurate as possible based on the daily feedback loop we have in place.