Cite as: 594 U. S. ____ (2021) 31 Opinion of the Court it off at any time within 27 days of an election. Arizona also makes special provision for certain groups of voters who are unable to use the early voting system. Every county must establish a special election board to serve voters who are “confined as the result of a continuing illness or physical disability,” are unable to go to the polls on election day, and do not wish to cast an early vote by mail. Ariz. Rev. Stat. Ann. §16–549(C) (Cum. Supp. 2020). At the request of a voter in this group, the board will deliver a ballot in person and return it on the voter’s behalf. §§16–549(C), (E). Arizona law also requires employers to give employees time off to vote when they are otherwise scheduled to work certain shifts on election day. §16–402 (2015). The plaintiffs were unable to provide statistical evidence showing that HB 2023 had a disparate impact on minority voters. Instead, they called witnesses who testified that third-party ballot collection tends to be used most heavily in disadvantaged communities and that minorities in Arizona—especially Native Americans—are disproportionately disadvantaged. 329 F. Supp. 3d, at 868, 870. But from that evidence the District Court could conclude only that prior to HB 2023’s enactment, “minorities generically were more likely than non-minorities to return their early ballots with the assistance of third parties.” Id., at 870. How much more, the court could not say from the record. Ibid. Neither can we. And without more concrete evidence, we cannot conclude that HB 2023 results in less opportunity to participate in the political process. 19 —————— 19 Not one to let the absence of a key finding get in the way, the dissent concludes from its own review of the evidence that HB 2023 “prevents many Native Americans from making effective use of one of the principal means of voting in Arizona,” and that “[w]hat is an inconsequential burden for others is for these citizens a severe hardship.” Post, at 38. What is missing from those statements is any evidence about the actual size of the disparity. (For that matter, by the time the dissent gets around to assessing HB 2023, it appears to have lost its zeal for statistical significance, which is nowhere to be seen. See post, at 35–40, and n. 13.) The