When the Supreme Court over the weekend rejected a petition to stop a Texas voter ID law from going into effect for the midterms, the left commenced its wailing and gnashing of teeth.

In her dissent, Justice Ruth Bader Ginsburg called the law “purposely discriminatory,” and everyone piled in behind her with denunciations of the Lone Star State’s blatant racism.

For the left, voter ID is tantamount to a poll tax. It is meant to suppress minority voters and is a last-gasp, unconstitutional scheme by the Republican Party to save itself by decisively shaping the electorate to its advantage.

If all of this is true, the nation is awash in neo-segregationist election rules. According to a recent Government Accountability Office report on voter ID laws, 33 states now have them, although the rules vary.

A valid ID is a necessity of modern life and requiring one to vote hardly seems an undue imposition. Especially if you are willing to give one out gratis. Of the 17 states that have strict requirements for a photo or government-issued ID, the GAO notes, 16 provide a free ID to eligible voters.

The critics reject this as yet another poll tax because people may not have the relevant underlying documents to get the free ID and there is a cost to obtaining them.

Well, yes. In Indiana, for instance, it costs $10 to obtain a birth certificate. In Arkansas, it costs $12. In North Dakota, $7.

The hard numbers suggest that the number of voters getting locked out by voter ID laws is diminishingly small. The GAO report focuses on the voter ID states of Kansas and Tennessee. In these states, voters whose eligibility to vote is in doubt may vote provisionally. Then, they have a period after the election to prove that they were indeed eligible and have their votes counted.

How many voters are showing up to vote, only to realize that they have been denied their rights by the ID requirement?

According to the GAO, in Kansas in 2012, 1,115,281 ballots were cast. There were 38,865 provisional ballots, and of these, 838 were cast for voter ID reasons.

In Tennessee, 2,480,182 ballots were cast. There were 7,089 provisional ballots and of these, 673 were cast for voter ID reasons.

In both states, about 30 percent of these voter ID-related provisional ballots were ultimately accepted. That means in Kansas and Tennessee, altogether about 1,000 ballots weren’t counted (and perhaps many of them for good reason), out of roughly 3.5 million cast. There you have it ladies and gentlemen, voter suppression! It is of such stuff that Jim Crow was made.

Indeed, voter ID is a scheme to suppress minority votes that is so nefarious that its effect can’t reliably be detected by the tools of social science.

The studies are all over the map and shot through with questionable assumptions and methods. The GAO looked at 10 studies on the effect of voter ID on turnout in general. Half of them found no effect. Four of them found it slightly suppressed turnout, and one showed it increased turnout.

Of the five studies, three found statistically significant effects on minorities, two did not. As a study last December in Political Research Quarterly notes, the idea that voter ID suppresses minority turnout “is strongly suggested in political discourse but lacks a strong empirical basis.”

The analysis by the authors of the Political Research Quarterly study concluded that “more stringent ID requirements for voting have no deterring effect on individual turnout across different racial and ethnic groups.”

For its part, the GAO used “a quasi-experimental analysis” to find that voter ID suppressed turnout in Kansas and Tennessee from 2008 to 2012, and that the effect was larger among African-American voters, but not Asian-Americans or Hispanics.

The states dispute the methodology. Kansas Secretary of State Kris Kobach argues that it makes no sense to compare turnout from 2008 with 2012 because in 2012 there wasn’t a Senate race in the state and therefore no get-out-the-vote effort. He points out that in 2000, when there also was no Senate race, turnout was 66.7 percent. In 2012, after the passage of the voter ID law, it was almost exactly the same, 66.8 percent.

Where you come down on this issue really depends on whether you think it’s reasonable to require the minimal effort to establish your identity of producing an ID at the ballot box or not.

The critics say that in-person voter fraud is extremely rare, although that is not an argument for leaving the system completely open to it. As my colleague John Fund points out, the New York City Department of Investigation last year had undercover agents try to vote as persons who were in jail, had moved, or were dead. They were successful 61 out of 63 times.

Is voting so important that it shouldn’t be tethered to an ID requirement? That’s not how we treat other important rights, as Hans von Spakovsky of the Heritage Foundation notes.

It takes an ID to buy a gun, a constitutional right. It takes an ID to get a marriage license or check into a hotel. No one goes around complaining that these requirements infringe on the rights of minorities to own a firearm, get married, or avail themselves of public accommodations.

Voting is inevitably going to require, even in the most latitudinarian states, some effort. You have to, at least most of the time, go to the polling place. You have to fill in the bubbles correctly. You have to deposit your ballot in a box. Not all people will go to the trouble to do this, or to do it correctly, which doesn’t mean they are disenfranchised.

The irony is that unhinged complaints about voter ID are, in this supposedly troubling new era of the poll tax, a turnout tool.