Early voting in person continues across Texas, and turnout continues to be at a record-shattering pace. Just over 1.4M had cast ballots in person and by mail in the 15 counties with the most registered voters through Wednesday, and turnout was steady in the counties from which we received reports on Thursday.

However, it is too early to conclude whether this early momentum can sustain itself, result in record overall turnout or change the electoral course the state has been on for more than two decades.

One can in fact see just about whatever pattern one is looking for in consistently higher numbers. Democratic activists believe increased turnout is a sign of a blue wave of voters that make the state a battleground, now and in the future. Republican activists see the increased turnout as voters angry with President Obama and the course of the nation turning out to vote down Hillary Clinton.

Partisans and analysts have espoused various reasons for record-smashing turnout through a handful of days of early voting, but the explanation may be far simpler and less transformative. Voters who would normally vote in a presidential general election may just be voting earlier than usual, and they’re being joined by new voters in numbers proportional to the number of first-time voter registrations.

Early voters have accounted for more than two thirds of all votes cast in the last two presidential elections in the 15 counties with the most registered voters, up considerably from 2004. The first six days of early voting – in person and by mail – accounted for 22% of all votes cast in 2004, 32% in 2008 and 34% in 2012. Just over 1M voted during those first six days in 2004, and that number rose dramatically in 2008 to 1.65M. In 2012, it rose slightly again to 1.76M. We have likely already eclipsed that number so far this year.

Republican political consultant Derek Ryan’s analysis of who is voting early shows that voters with Republican primary voting history (36%) slightly outpace the number of voters with Democratic primary voting history (30%). Based on 2016 primary turnout, one would expect to see a much greater difference, because there were nearly twice as many Republican primary voters as Democratic primary voters.

One of the reasons for this is ballot-by-mail voters. Ryan found that 35% of voters casting ballots by mail have a Democratic primary history, which is closer to the 40% with a Republican history than in the 2012 election, when the gap between the ballot-by-mail voters with primary history was more like 24 points.

That said, the fact that such a high percentage of early voters has recent primary history suggests that the bulk of early votes so far have been cast by partisans – likely straight-ticket voters – which will leave the bulk of later-day (and Election Day) turnout to voters who are less committed to a particular party’s candidates, as a group, than the voters whose ballots are already in the box.

In 2008, the number of primary voters (Republican and Democrat) was 53% of the total number of general election voters. Some primary voters no doubt did not return for the general election for various reasons, but it’s a good bet that at least half of the general election voters also voted in a primary that year. The combined number of primary voters this year slightly exceeded 2008. If turnout is around 8M for this general election, then primary voters would represent around 53% of general election voters.

Ryan’s analysis indicates that about 10% of early voters have no prior recent primary or general election voting history. That is in line with the proportional increase in the number of new non-suspense voters registering between the March primary (12.7M) and now (13.4M). Non-suspense voters are a better yardstick because those voters’ addresses are believed to be accurate and they have either recent voter history or a recent new registration.

Hispanic/Latino-majority counties are showing some of the higher percentages in new voter participation. Voters with no recent prior history comprise 15% of the early vote in Hidalgo Co. and 14% in El Paso Co., the two highest such figures in Ryan’s analysis. However, there are significantly more new voters in more traditionally Republican counties, even though they represent a smaller percentage of votes cast so far.

Highlights of early voting turnout on Day Four (Thursday):

  • In Harris Co., slightly more voters cast ballots in person on Thursday than in any other day so far. Overall, more than 366K have voted early through four days. It took six days to reach that number in 2012, eight days in 2008 and 11 days in 2004.
  • In Tarrant Co., 40K voted in person on Thursday, the lowest number so far but just 11% off the single-day high. Nearly 196K have voted early through four days, 33% more than through four days in 2012, 41% than 2008 and 126% more than 2004.
  • In Bexar Co., more than 39K voted in person on Thursday, just below Wednesday’s high point.
  • In Travis Co., nearly 33K voted in person on Thursday, marking the second consecutive day with slightly declining turnout. Overall, the number of early votes cast in Travis Co. through four days is more than double the 2012 pace. In that year, it took nine days to reach the four-day total of this year (154K).

Early voting got off to record-setting starts in 2012 and 2014 only to fall short of those years. We are far enough ahead this year that a record early voting turnout is likely. It is a jump in magnitude similar to 2008. However, in 2008, fewer people voted on Election Day in the 15 big counties than in 2006, a gubernatorial election year where turnout overall was barely half of what it was two years later.

A significant drop in Election Day turnout would support two theories. First, it would lend evidence to the idea that voters are simply voting earlier than in previous years. Second, it would bolster the idea that the true partisans’ votes are coming in early, leaving Election Day to far less committed voters, a significant number of which stay home.