Analytics

Understanding Absentee and Early Voter Trends in North Carolina

By Charles Zhu - November 2, 2020

Across the country, states are reporting unprecedented levels of early and mail-in voting. This change in voter behavior is likely to have a significant impact on the overall result as well as how polls, predictions, and coverage are carried out in future years.

In elections, local details matter, so with a few days to go before the counting begins in earnest, we helped The Raleigh News & Observer dig deeper into the details in a state many are saying will play a key role in the outcome: North Carolina.

Early voting for North Carolina ended on Saturday, October 31st, and as of November 1st, more than 4.5 million North Carolinians have already cast their ballot. According to The News & Observer, “that’s more than 95% of all the N.C. voters who cast ballots in 2016.” Of these early votes, more than 3.6 million were cast in person, and 1,009,880 have voted absentee.

This year, North Carolina took an unprecedented step in transparency by publishing ballot-level data on both in-person and absentee (mail-in) ballots. Using Sisu, we looked through millions of factor combinations to surface any trends in the data and shared our findings with the News & Observer.

Demographic differences in rejection rates

The first observation Sisu surfaced highlighted some big disparities in ballot rejection rates. Across all the early ballots statewide, 0.798% have been rejected by the end of October. Absentee ballots have a rejection rate of 3%, whereas in-person is just 0.24%. These rejection rates may seem small, but in North Carolina smaller margins have been a deciding factor in past elections. As the News & Observer reports,

“In 2016, the governor’s race came down to about 0.2% of the vote. And the presidential races in North Carolina in 2008, 2012 and 2016 were all decided by between 0.3% and 3.7% of the vote.”

Digging in deeper, we found that in this year’s early voting, the ballot rejection rate for mail-in ballots is unevenly distributed. Let’s take a closer look:

In North Carolina, 5.4% of all mail-in ballots from Black voters were rejected

While the total rejection rate for absentee ballots is 3%, absentee ballots from self-identified Black voters are being rejected at a rate of 5.3%, 2.1x higher than the rest of the population. As The Raleigh News and Observer reported,

“State officials say ballots are rejected for missing things such as witness signatures, which could be the result of people being unfamiliar with the requirements. Ballots or the envelopes they come in have no indication of race, they said.”

As of October 30th, Black voters have cast 15.2% of all ballots in North Carolina, and of those 84% are from registered Democrats. We looked to see if there were other trends around race and ballot rejection and also found that among absentee ballots cast by Latinx voters, the rejection rate was 4.3%. That’s almost 2% more than the rejection rate of white votes, which stands at 2.4%. The latino vote is 2.4% of the vote so far.

In a few key counties, rejection rate of ballots from Black voters is >10%

As we drilled in further, we found even more extreme disparities in the rate of rejection in a few key counties. For example, in some counties (like Brunswick, Vance, and New Hanover) Black-submitted mail-in ballots are rejected at a rate greater than 10%. In Beaufort County, the rejection rate for Black-submitted was 15.5%, while the rejection rate for Native American-submitted ballots in Robeson county was over 12.5%.

This increased rate of rejection underscores that discrepancies in voting behavior are inherently local. According to the North Carolina State Board of Elections, mail-in ballot completeness is assessed by the county board of elections, which could explain the wide deviance between a few counties and the rest of the state.

What’s driving these increases in ballot rejection for Black voters?

With ballots from Black voters being rejected by 5.4% overall, we wanted to look further at the underlying why. Fortunately, North Carolina’s data is very granular and includes a lot of useful detail. Of the 8,317 rejected ballots from Black voters in the dataset, the cited reason for rejections are:

  • 32.6% are “spoiled”
  • 26.5% are awaiting a cure
  • 24.8% have witness info as incomplete
  • 11.8% have been accepted after cure
  • 3.4% are returned undeliverable

Even for experienced voters, the steps for correctly submitting and filling out a mail-in ballot can cause confusion that leads to a rejected or “spoiled” vote. Absentee ballots require steps most voters don’t often do when voting in person — like signing the envelope and having a witness who provides an address and signature.

The witness signature has been especially problematic in North Carolina, and appears to disproportionately invalidate Black mail-in ballots. A witness signature is a third party signature that attest that the voter, not someone else, signed the ballot. Previously, state officials had allowed voters to fix the lack of a witness signature by returning an affidavit that it was their ballot, but as of October 14th a federal judge ruled that preventing absentee voters from fixing their ballot if it lacked this third party signature, according to Reuters.

Thankfully, this doesn’t mean these votes will never be counted. According to the North Carolina State Board of Elections, county boards are required to contact any voter with any rejected ballot within one business day, and voters have until November 12th to “cure” their vote.

Watching the data on Election Night and beyond

According to the News & Observer, “The State Board of Elections said Sunday it expects 97% or more of all ballots cast in North Carolina will be counted and reported Tuesday night.” Unlike other states, election officials are allowed to count mail-in votes as soon as they arrive.

As votes are counted and cured, we’ll be watching the data to see how these rejections trends changed or are addressed. If this level of analysis is helpful for your reporting, reach out to our team.


Read more

Analyzing Racial and Geographic Disparities in COVID-19 Cases with Snowflake Data Shares and Sisu

Using data from Snowflake Data Shares and Sisu, our team took a look at the racial and geographic trends in new COVID-19 cases. Take a look at the differences between the rise of cases in March to the most recent surge of cases in July.

Read more

Four Reasons Why Now is the Time for Augmented Analytics

Insights from Sisu customers on augmented analytics and how adopting this proactive approach lets analytics teams fully unlock the value of their data.

Read more