What Can NC’s Early Voting Data Tell Us (and What They Don’t)

By Michael Bitzer and Christopher Cooper 

In person-early voting kicked off yesterday in North Carolina, but before a single in-person vote was cast, more than 65,000 votes had already been accepted across all 100 counties. 

A host of political scientists, political analysis, and political consultants will be reporting patterns from these early voting numbers. Look for posts from the Kevin Bacon of North Carolina politics, Gerry Cohen (no one in #ncpol is more than six degrees removed), as well as from The John Locke Foundation’s Andy Jackson, and University of Florida Political Scientist Michael McDonald. We will (individually and collectively) probably throw some of these results into the mix, too. 

So, before we get to the season of early voting data, it’s worth thinking through the mail and early voting process–why we do it, what we can learn, and what we can’t. 

Why Did We Have Votes Accepted Before Early voting Even Started?

Let’s start with why ballots are being accepted now. Until September 24 when mail ballots were mailed to regular citizens, all of the ballots cast thus far were made possible not through North Carolina law, but rather through the Uniformed and Overseas Citizens Absentee Voting Act (UOCAVA). 

Whereas most of the rules around elections in the United States are determined at the state level, the federal government decided that we needed a standardized set of guidelines for military voters who are stationed overseas. This makes sense, right? If you’re stationed in Antarctica, ambling down to the local polling place on election might prove difficult. So, this law, signed by President Reagan in 1986, is supposed to take the unique situations of these voters into account by setting federal deadlines for when their ballots are sent out and how they are returned (see here for a great explainer from the Congressional Research Service). 

From September 24 through October 16, regular citizens who had requested a mail ballot could also return a mail ballot (either through the mail or in person to their county board of elections). 

So, What Early Voting Data Tell Us and Why is it Worth Paying Attention?

There is a lot of misinformation swirling around about election administration and election security. Following patterns in mail and early voting is a great way to combat this misinformation. You can follow, for example, whether the number of ballots returned exceeds the number requested (it won’t), who requests and returns each mail ballot, and whether people cast more than one ballot. 

Following these data also give us a guide to what we may see on election night. If, for example, we see that Democratic voters are more likely to participate in early voting, then we know that the first dump of election results on election night will likely look much better for the Democratic Party than the final tallies (the so-called “blue mirage”). Following these patterns, therefore, can help right-size expectations.

These patterns are also important for what they tell us about how North Carolinians are casting their votes. The last few cycles have seen an increasing partisan divide on method of voting–with Democrats increasingly favoring early and mail voting and Republicans favoring election day voting. Republican groups like American Majority (headed in NC by Dallas Woodhouse) have attempted to combat this and encourage Republicans to “bank their vote.” Will those efforts be successful? Following the early voting data will help answer that question.

We can also evaluate a host of other questions in real-time using these data including: How many people are taking advantage of same day registration (SDR)? Do SDR voters tend to be younger? From out of state? Are they from certain areas? How quickly are mail ballots “cured” in North Carolina? Do we see patterns in whose ballots are not counted because of various deficiencies? Do the patterns in early voting use follow what the polls are predicting? 

In other words: a lot of questions, but we need to let the data tell us (if possible) any of the answers.

Early Voting Data Are Not Predictive of Final Election Outcomes

There is often a tendency to try to use patterns in early voting data as an early sign of who might win or lose the election. Please don’t do this. I mean, you could, but it would be like bringing a fishing pole to a home run derby–it’s just the wrong tool for the job.

To give folks a sense of things from 2020 (and yes, 2020 was an election like no other), the breakdowns by party registrations for each vote method would be the following, starting with absentee by mail ballots.

Now, again, a cautionary note: if we see 2024’s absentee by mail ballots equal the number cast in 2020, then both of us will be out of a job real soon. But note the Democratic and Unaffiliated dominance of this vote method, resulting in a landslide victory (by mail) for Biden (getting 70 percent) to Trump’s 28 percent in just this vote method.

Turn to absentee in-person voting, with the following party registration breakdown:

Data from NC State Board of Elections Voter History File; compiled by Michael Bitzer

Since 2008, registered Republicans have become more and more accustomed to voting ‘early’ in-person and are now an even division of this vote method with registered Democrats. Trump edged out Biden in this vote method, 52 to 47 percent. 

Finally, look at Election Day’s vote and the party registration breakdown in that method:

Data from NC State Board of Elections Voter History File; compiled by Michael Bitzer

Registered Republicans have their advantage in this dynamic, and it showed in the results for this vote method: Trump won 2-1 over Biden.

But when all three major vote methods are pulled together, here’s the total electorate by party registration, resulting in the 49.9 to 48.6 percent win for Trump. 

Data from NC State Board of Elections Voter History File; compiled by Michael Bitzer

The overall electorate has a slight registered Republican advantage compared to their registration percentage (see this post for a discussion on what might be NC’s potential electorate), but there’s a fairly evenly division among the party registrations. 

As a reminder, here are the election results for each vote method used in 2020’s North Carolina general election:

For this year, let’s say that Democratic turnout is leading Republican turnout by a greater margin than in previous elections. That could mean that Democratic turnout is up (and therefore Democrats are more likely to win competitive elections), or, it could mean that Democrats are increasingly likely to vote early instead of on election day, whereas the opposite trend is true for Republicans. There’s just no way to know until election day. 

So, pay attention to early voting data for the insights they provide into our electoral process and how election administration works. But don’t think they’re going to help you predict who will win the election. If that’s your goal, you’re better off practicing aleuromancy than relying on patterns in early voting data.

____

Dr. Michael Bitzer holds the Leonard Chair of Political Science and director of the Center for North Carolina Politics & Public Service at Catawba College. Chris Cooper is Madison Distinguished Professor of Political Science and Public Affairs and Director of the Haire Institute for Public Policy at Western Carolina University. Both Bitzer and Cooper will spend an inordinate of time analyzing NCs mail and early voting data from now until the election day.