Brian Arbour: What early voting numbers do and do not mean
It’s 2020, and everything is different. Some things are a little different, and some things are a lot different.
One of the things that is a lot different this year is early voting. More people are doing it than in previous years. A lot more.
The U.S. Elections Project reports that over 92 million ballots have already been received by election officials, as of Sunday morning. This not only outpaces the 47 million early votes recorded in 2016, but is nearly two-thirds of the nearly 139 million votes cast in that election. Leading the way in early voting is Texas; more votes have been cast early in the Lone Star State in 2020 than were cast in the entire 2016 election.
I am one of those 92 million. The state of New Jersey sent ballots by mail to every registered voter in the state. I took my ballot to our town hall, and let my 8-year old son have the honors of putting it in the dropbox.
The increase in early voting this year has led to much speculation on what these results will mean for election night. Does one candidate or party have an advantage due to the geographic, racial, age or partisan composition of early voters?
But my voting story helps to explain why these efforts are futile. Many of those who are voting early this year are regular voters who are choosing to vote early because, in response to the coronavirus pandemic, states have expanded early voting options to reduce crowds on Election Day.
As a result, we have less information about who is going to vote on Election Day itself, and who it will benefit. Will it be a strong or a weak turnout? Will Election Day turnout come more from one group or another? We can estimate and make predictions, but we do not know.
While 2020 has compelled changes large and small in different areas of American life, one area where it has not produced changes is in attempts to divine clues about election results from early voting data. Pundits on editorial pages and in social media try to parse through the early voting numbers to identify advantages for one side or the other. These rarely turn out to be correct.
The changes 2020 has wrought on election administration has made it more difficult, if not impossible, to read the tea leaves provided by early voting data. There are two reasons for this.
First, the increase in early voting means that there is no historical pattern of voting behavior that one can use to assess this year’s early vote. And second, polling data indicates that there is a great partisan difference in whether voters are going to vote early or on Election Day. These facts make it nearly impossible to determine what advantages one candidate has in early, or in Election Day voting.
As discussed, I usually vote on Election Day but was able to vote early this year. This is is very common; political science research and election analysis have long found that early voting usually cannibalizes Election Day voting. Certainly, some of those who have voted already are new or sporadic voters who did not vote in 2016. But the vast majority of those who have voted are regular voters, even if they usually vote on Election Day.
Thus, more early votes in 2020 mean fewer Election Day voters, especially among those who are regular voters. Who is left to vote on Election Day? It will not be the usual set of people. Because both early voters and Election Day voters are different from the usual sets who select each option, we lack the ability to compare each group to subsets of voters from 2016, or any other previous election.
It is difficult to parse advantages in early voting data even under the best of circumstances.
If early voters in 2020 are younger or more urban than those who voted in 2016, we don’t really know if that means the entire electorate will be younger or more urban than previous electorates, or if more of those voters chose to vote early in 2020 than in previous years.
Early vote analysis this year is also confounded by partisan differences in when to vote. This year, the most committed Republicans seem to be following the cues provided by the nation’s top Republican – President Donald Trump – that early voting by mail is “unfair,” a “disaster” and a “scam.” Poll results and early voting data show that Democrats favor voting by mail, which is the most common form of early voting being used this year.
Our Fox News polls show the partisan dichotomy on how to vote. Our most recent set of polls asked voters in four Rust Belt states if they would vote early or if they would vote in person on Election Day. In the swing state of Wisconsin, those who have or are planning on voting early favored Biden 65-28. Those who will cast their votes in person on Election Day favor President Trump 37-56. Our polls showed similar results in Michigan, Ohio and Pennsylvania.
We see more Democrats voting early, and we will see more Republicans show up at the polls next Tuesday. Will the Republican Election Day surge match the early voting efforts of Democrats? We do not know now, and will not know the answer until next Tuesday.
So what should you do if you read an article, view a social media post, or hear chatter from a friend or family member about how some bit of early voting data portends victory or doom for your favored candidate? Take it with a grain – nay, a pillar – of salt.
It is difficult to parse advantages in early voting data even under the best of circumstances. Trying to do it in 2020 is near impossible. We will all have to wait until actual results are released on election night to know what the early voting numbers – and the Election Day numbers – mean.