ATLANTA (AP) — Georgia's highest court on Thursday affirmed a lower court dismissal of a lawsuit challenging the outcome of last year's race for lieutenant governor in a case that put a spotlight on the outdated voting machines the state is in the process of replacing.

The lawsuit alleged that an undercount of tens of thousands of votes in the lieutenant governor's race was likely caused by problems with the state's paperless touchscreen voting machines that either caused voters not to vote in that race or those votes to go uncounted.

That assertion is "wholly unsupported" by the record in the case, so the trial court wasn't wrong to conclude that the plaintiffs "failed to meet their burden of showing an irregularity in Georgia's electronic voting system sufficient to cast doubt on the 2018 election," Georgia Supreme Court Justice Sarah Warren wrote in the unanimous opinion.

Republican Geoff Duncan beat Democrat Sarah Riggs Amico by 123,172 votes to become lieutenant governor. Amico is not a party to the lawsuit, which was filed in November by the Coalition for Good Governance, an election integrity advocacy organization; Smythe Duval, who ran for secretary of state as a Libertarian; and two Georgia voters. It was filed against Duncan and election officials.

Senior Superior Court Judge Adele Grubbs dismissed the lawsuit in January. In their appeal to the high court, the plaintiffs argued that Grubbs erred by not allowing discovery prior to trial.

"We conclude that, given the Election Code's statutory mandates and the broad discretion trial courts are given to manage pretrial discovery, the trial court did not abuse its discretion," Warren wrote.

Almost all voters make a selection for the race at the top of the ballot — the governor's race in this case — and then numbers drops slightly for down-ballot races, the appeal says. But in November, the drop in total votes recorded in the lieutenant governor's race compared with the governor's race was much bigger than normal, and other down-ballot races didn't experience a similar decline.

The plaintiffs contended that, based on historical voting patterns, there may have been an undercount of more than 127,000 votes in November's election.

As part of the discovery process, plaintiffs sought access to an electronic state database containing election data as well as access to the internal memory of voting machines so their experts could do a forensic examination. They argued that a forensic examination was the only way to tell whether the undervote was caused by a programming defect.

The access the plaintiffs sought raised security concerns, and the secretary of state and county election officials provided an alternative way for the plaintiffs to get the information they sought but they rejected that offer, argued a lawyer for Secretary of State Brad Raffensperger, who had been dismissed as a defendant but had filed a brief in support of the lower court's dismissal of the lawsuit.

The Coalition for Good Governance had been dismissed as a plaintiff, but executive director Marilyn Marks remained invested in the case. She said she was surprised by the court's ruling, saying it allows the government to conduct elections on electronic machines with no way for citizens to hold them accountable because they can't access the evidence.

"What Georgia is saying is, 'No, this is a function to be conducted by the government and all the secrets will be kept by the government,'" she said.

The state is in the process of implementing a new system that includes touchscreen voting machines that produce a paper record, including a human-readable summary of the voter's selections and a machine-readable code used by a scanner to tally the votes.

Critics say the new machines have many of the same security problems as the old ones.

They also note that while although printed paper record includes a human-readable summary of the voter's selections, the scanner tallies the votes based on a machine-readable code. That's problematic because voters can't be sure that the code on the paper accurately reflects their selections and meaningful audits can't be done, they argue.