Put Not Your Trust In Early Vote Counts

[ Posted Monday, November 5th, 2018 – 16:59 UTC ]

It happens every time we have an election. The numbers-crunching uber-wonks among us have to work themselves into a frenzy for Election Night, because that's when the data will be coming in fast and furious, so they've got to be at their peak performance. But, unfortunately, before the polls actually close and the numbers start flowing, there is a dearth of data in the buildup period. There are last-minute public opinion polls to watch, but they don't say much that is all that different than other pre-election polling. So the temptation is to take the only hard numbers available and overanalyze them to death. But my experience causes me to issue the following warning to all who are obsessing over the midterm elections: Put not your trust in early vote counts.

Until the votes start to actually be tabulated, these are the only numbers made public by the people running the elections in each state. But early voting numbers are just not determinative of much of anything, even though all the wonks are having lots of fun making all sorts of predictions based upon these numbers. But when you get right down to it, the only thing early voting numbers really tell us is how many people voted early, period. Which, in and of itself, doesn't really tell us a thing about the likely outcome.

The first thing to consider about early voting is that there are two long-term historical trends that have been happening, both of which have been increasing the number of votes cast early, over time. Most people are either unaware of it (or just don't think about it much), but there are fully 50-plus sets of rules in place for how votes are cast. The states set these rules individually. They wildly differ from state to state, but most people just generally assume that voting takes place everywhere pretty much the same way as it does in their state. This is wrong, because there are a lot of major differences, state to state.

One of these big differences is how absentee voting takes place. Some states are unbelievably restrictive, forcing the voter to actually provide proof that they will not be present in their precinct on Election Day for a valid reason. In other states, all you have to do is request an absentee ballot and one is automatically sent to you. You don't have to provide a reason for absentee voting other than "I want to vote using this method," in other words. And in a handful of places, everyone gets their ballot in the mail. Over time, this has been moving in a less-restrictive direction. More and more states have been making it easier to get absentee ballots in the past few decades. Some places have countered this trend, but overall the movement is towards more absentee ballots being provided.

Oregon is one of the states that have moved completely to a "vote by mail" system. Everyone is automatically registered to vote when interacting with state agencies (like when you get a driver's license) and then every registered voter gets a ballot in the mail for every election. You mark up your ballot, sign it (or the envelope), and send it back in. There are dropoff collection boxes provided so people who don't trust the postal system can vote early, and on Election Day there are no actual "polling places," just dropoff sites (for those who still like the tradition of voting on Election Day itself). This system has two very big things going for it: it's cheaper than the alternative, and it works better, too. More people vote using such a universal vote-by-mail system than would otherwise. Turnout improves, and the state saves money at the same time -- what's not to love? Over time, more states can be expected to adopt this system.

The other big-picture trend at play is the introduction and expansion of early voting itself. Even in states with restrictive absentee voting, in-person voting happens days (sometimes weeks) before Election Day. Again, some individual states have tried to pull back from early voting, but for the most part the trend is towards allowing people to vote early if they choose to. And again, over time, more and more people are deciding to vote early so they don't have to worry about it on Election Day itself. It's just more convenient for many to vote on a weekend, a few weeks before Election Day, than having to fit it into a busy Tuesday.

Both of these long-term trends have the same result: more and more people voting early. So when you hear that "a record number of people have voted early" in an election cycle, it's shouldn't be all that surprising (or notable), because that's where the long-term trend has been headed. But this often gets lost in the breathless reporting on the early vote numbers.

But the biggest error the over-analyzers of early voting make is that the number of people casting early ballots really tells you almost nothing about what the outcome will be. At most, it may signify overall enthusiasm for a particular election, but this doesn't usually translate to partisan advantage.

There's one big question that is never really asked: if more and more people vote early, will that mean more people voting on Election Day, or fewer people voting on Election Day? It can be either one -- if overall enthusiasm is through the roof for an election, then you'd expect both the early vote numbers and the in-person numbers on Election Day to be higher than normal. But, over time, if more and more people decide it is just easier to cast their ballot early and they start regularly doing it in every election they vote in, then eventually that will mean that fewer and fewer of them show up on Election Day (because most of them will have already voted).

One question I'd dearly love some superwonk to do a deep dive into is how increased early voting affects the way the results are reported on Election Night. This would probably hinge on a basic question: do the precincts and counties count the early votes first (ahead of time, in other words), or do they just throw them all in a pile and count them at the end of the night? That might be meaningful for late swings in the vote totals as the results trickle in, one way or the other. Which is why I'd love it if someone did a thorough analysis.

Early voting totals may show increased overall enthusiasm about voting, but this also doesn't necessarily translate into clear partisan trends. I've seen Democrats (and Republicans, to be fair) make this mistake before. They read the early voting numbers, microanalyze them, and decide that they're going to win a big victory. However, on Election Day, the count is either closer than anticipated or that predicted victory turns into a defeat. Just because lots of people vote early, it doesn't necessarily mean that lots of your voters voted early. And even if it does, that doesn't necessarily predict what the in-person turnout on Election Day is going to be.

Take just one demographic -- young voters. Let's say there is a boom in young early votes cast. Democrats, naturally, read into this a big advantage. However, because early voting is so wildly popular among young people -- people who don't see any point to the traditional ceremony of casting your vote in person on Election Day -- it means that more young voters cast their ballots early than vote on Election Day. What this means on Election Day is that the people who do turn out to vote in person will be older, for the most part. Which can give Republicans the advantage on Election Day -- which could completely counter any advantage Democrats built up through early voting.

Also, since the votes are not counted yet, it really is impossible to tell who the early voters have voted for. Some states report their early voting totals broken down into party preference -- this many Republican-registered early votes, and this many Democratic-registered early votes. But party registration is no guarantee of how that individual person voted. How many registered Democrats voted for Trump, after all?

In the end, the only thing that early voting totals tell us is really only of interest to political scientists tracking long-term trends: what is the electorate's preferred method of voting? This is a function of each state's voting laws, of course. How easy is it to get an absentee ballot and mail it in early? How many in-person voting sites and early voting days does the state offer? What percentage of the total votes cast are cast early? How does this percentage change, over time? As I said, this is of interest to those who study such long-term trends, but each individual data point (each individual election) doesn't tell us much at all.

So put not your trust in early voting data. If anything, such numbers may show an increase (or decrease) in public enthusiasm for that particular election. But just because voters from one side are enthused doesn't mean that the voters on the other side aren't equally enthused. Just because a lot of your voters voted early doesn't mean that the other side won't make up the gap on Election Day. These are really the crucial things to keep in mind when listening to others make wild predictions based only on how many people voted early.

-- Chris Weigant


Follow Chris on Twitter: @ChrisWeigant


14 Comments on “Put Not Your Trust In Early Vote Counts”

  1. [1] 
    nypoet22 wrote:

    you fooled me CW, you were quoting psalms and i thought it was yoda. anyhow, with only a few hours until the election you could still write a column about pie. it wouldn't be as good as if you'd lived up to your "reality-based" moniker sooner, but it would still be pie.


  2. [2] 
    Kick wrote:

    I hear you, CW, but FYI: Jon Ralston in Nevada informs that Democrats have built up such an early vote lead there that he predicts a good night for Democrats. Ralston says if the first results from Clark show double digit leads, then the math from the rural counties is not going to be enough to save Heller. He's predicting... based on math and not polls... a win by Jacky Rosen.

    I've never known Ralston to be wrong, but there's obviously a first time for everything. :)

  3. [3] 
    TheStig wrote:

    Russia wii be very unhappy if all states follow Oregon and " go postal" Much less vulnerable to cyber attack.

    I voted 3 weeks ago.

  4. [4] 
    LeaningBlue wrote:

    Here's a knee-slapper. But first the background. The RCP Generic Ballot polls post 10/24 are (all +Dems, oldest first): 7,9,9,5,9,8,13,7. Except for one more, this morning. Wait for it.

    Rassmussen : Republicans +1.

    That was good enough for Zero Hedge to trumpet the headline Stocks Surge After Latest Poll Shows GOP Retaining The House.

    Their justification for believing the Rassmus bullshit? "Rasmussen was the only major pollster in 2016 to predict a Trump victory."

    Okay, fair enough, except for one thing. They were polling the national popular vote in that poll.

    CW is right. Believe nothing and no one. Go out to dinner. Take in a movie. Watch the Santos-Vinnick election West Wing episode on Netflix. Check back when CA is being counted, if then.

  5. [5] 
    goode trickle wrote:

    So... this is the first election I will not be able to vote in quite some time.

    Mostly due to the inadequacy of our absentee voting system and the fact I have not been back to the US to collect my ballot.

    My challenge to all of you activists is to get out and get 5 people to vote in my place. Tell them that it is their duty to step up and participate for those of us who can't.

    In the meantime I will continue to be out on a project with a big white ship that truly shows how great Americans can be.

  6. [6] 
    Kick wrote:


    Russia wii be very unhappy if all states follow Oregon and " go postal" Much less vulnerable to cyber attack.

    *laughs* All states going postal would be the Republicans' worst nightmare.

    I voted 3 weeks ago.

    I would wager it feels like 3 months ago. :)

  7. [7] 
    James T Canuck wrote:

    It was said best..."Muscled, black with steel green eye
    Swishing through the rye grass
    With thoughts of mouse and apple pie
    Tail balancing at half-mast
    ---And the mouse police never sleeps--
    Lying in the cherry tree
    Savage bed foot warmer
    Of purest feline ancestry
    Look out, little furry folk
    He's the all night working cat
    Eats but one in every ten
    Leaves the others on the mat
    --And the mouse police never sleeps---
    Waiting by the cellar door
    Window-box town crier
    Birth and death registrar
    With claws that rake a furrow red
    Licensed to mutilate
    From warm milk on a lazy day
    To dawn patrol on hungry hate
    No, the mouse police never sleeps
    Climbing on the ivy
    Windy roof top weathercock
    Warm blooded night on a cold tile.

    GTFO and vote.

    Happy voting day, my friends.


  8. [8] 
    Kick wrote:


    My challenge to all of you activists is to get out and get 5 people to vote in my place. Tell them that it is their duty to step up and participate for those of us who can't.

    Five people, you say? We had a goal of 500. At the end of early voting, my peeps and my loaned vehicle had collectively transported a count of 476 people to the polls. By the end of election day today, we expect to easily surpass that goal. There are lots of vehicles doing the same thing with the aim of flipping the heavily gerrymandered district of TX-32 from Red to Blue. It is no small order to flip any district in deep Red gerrymandered Texas.

    Texas will flip Blue eventually or Republicans will be forced to soften their far-right positions in order to maintain their hold on power. The fact that Republicans nationwide are promising to protect preexisting conditions and aren't running on killing the Affordable Care Act a.k.a. Obamacare speaks volumes.

    In the meantime I will continue to be out on a project with a big white ship that truly shows how great Americans can be.


  9. [9] 
    Paula wrote:

    [8] Kick: 476! So cool!

    Off to do one last canvass - in part to distract me while the hours pass. Getting anecdotal reports of high turnout in parts of Ohio.

    Hoping Sane and Decent America prevails.

    Good luck to all!

  10. [10] 
    John M wrote:

    [5] goode trickle

    "My challenge to all of you activists is to get out and get 5 people to vote in my place. Tell them that it is their duty to step up and participate for those of us who can't."

    I got 5 of 6 people I know to vote here in Florida. Hopefully number 6 will also. Hardly compares to Kick, but hey every little bit helps!

  11. [11] 
    Balthasar wrote:

    Paula [9] Getting anecdotal reports of high turnout in parts of Ohio

    Don't read too much into that, P: the Republican Sec of State of Ohio has a history of providing too few voting machines in poor areas and college towns, causing long lines to form.

  12. [12] 
    Balthasar wrote:

    In case you missed it: Randy Rainbow just posted another video titled 'Voting' - a parody of the song 'Loathing' from "Wicked". enjoy.

  13. [13] 
    neilm wrote:

    Interesting article on how right wing news organizations view their charter:

  14. [14] 
    ListenWhenYouHear wrote:

    I just wanted to thank all of you who got out and canvased your communities and encouraged folks to vote this year! It says a lot about you that you care enough to give of your own time for this country, and I think you deserve our gratitude. I tip my cap to ya!

    — Russ

Comments for this article are closed.