Voters line up outside the Vigo County Annex in Terre Haute, Ind., on Nov. 5 to take advantage of the final day of early voting. (Austen Leake/Tribune-Star/AP)

You’ve probably seen the data in various places (including The Post): There’s been a surge in early voting in a slew of states. In 35 states, as of Tuesday morning, more votes have been cast early than were cast early in 2014. In Texas, more votes have been cast early than were cast in 2014 overall. It’s a downright phenomenon!

But ... what does it mean? Does it mean a massive wave of enthusiasm? Or is it a continuation of a trend? Early voting has been trending upward in recent elections in the United States, as has the population. So shouldn’t we expect increases in early voting?

Two weeks ago, we looked at numbers from the Democratic data firm TargetSmart that offered some insight into this question. The firm breaks voters into four groups, depending on how often they usually vote. There is the fifth of the registered voter base that votes most frequently, dubbed “super voters.” The group that’s about in the midrange of voting propensity are “frequent voters.” People who have voted at least once but aren’t in the other categories are dubbed “infrequent,” and people who have never voted at all make up a category of their own.

If the early-voting surge were something that reflected unusual interest in the campaign, we might expect that the percentage of votes cast early by infrequent or new voters had increased. Early voting is, after all, just voting, and if the surge in early voting is just people who vote all the time simply voting at a different point in the cycle, that doesn’t tell us much about enthusiasm. If this election is spurring new interest, though, we would expect infrequent and new voters to make up more of that early-voting electorate.

Nationally, we see that happening. Compared with 2014, new and infrequent voters make up slightly more of the early voting electorate, while “super voters” make up slightly less.

(Philip Bump/The Washington Post)

More than three-quarters of the early vote, though, is still people who generally vote.

In some states, but not all, we see similarly modest increases in the density of infrequent voters. In the states in light gray below, new or infrequent voters made up about the same amount of the early-voting electorate as in 2014 or slightly less of it. States in dark gray saw increases in the density of infrequent voters relative to 2014. States indicated in red had growth of 50 percent or more in the density of infrequent voters.

(Philip Bump/The Washington Post)

It’s a mix! Included in the states with the biggest surges are some with the most closely watched races, notably Georgia and the aforementioned Texas. But there are also some states with surges in the density of infrequent voters that don’t have such significant races on the ballot, such as New York.

One important indicator: In no state did the density of infrequent voters this year top the density of infrequent voters seen in 2016 early voting.

If we compare the change in the density of infrequent voters since 2014 with the races that are on the ballot — using 270ToWin’s compilation of contested Senate and gubernatorial races and Cook Political Report’s House race ratings — we see that the biggest surges in infrequent early-voter density only occasionally overlap with important races. (The color coding is the same as above.)

(Philip Bump/The Washington Post)

In some places with big races, such as Florida, the increases have been much more subtle.

We must, of course, note that state at the bottom: Tennessee. Tennessee’s share of new and infrequent voters doubled in early voting compared with four years ago, an increase that you would not be criticized for noting overlaps with pop star Taylor Swift’s encouragement to her mostly young fans to register and vote. Granted, Tennessee also had one of the lowest densities of infrequent and new voters in 2014, so it’s easy to outpace that number.

So what’s the answer to the question of how big an electoral surge we should expect in 2018? Well, unfortunately we don’t yet have one. Once all the votes have been cast and counted, we’ll know how unusual the electorate was overall — and what these particular tea leaves had been trying to tell us.