In 17 days, Granite State voters will head to the polls to exercise what is, arguably, their most well-known right and duty: voting in the first-in-the-nation New Hampshire presidential primary. When they do, many will cast votes counted by machine. We hope all will go well. But there’s no guarantee. Absent a campaign-requested recount, there won’t be any verification of the results those machines offer. And that’s because the state’s highest elections officials refuse to allow it.

A handful of Monadnock Region residents has been sounding the alarm regarding the vulnerability of voting machines in New Hampshire for several years. And, as noted in a recent report by Sentinel staff writer Jake Lahut, the N.H. Secretary of State’s Office has been refusing to allow local polling officials to even conduct random cross checks by hand.

Such hand counts, or audits, could go a long way toward putting voters’ minds at ease regarding the efficacy of the machines upon which so much of our election infrastructure relies. It’s not just a worry for the conspiracy-minded; beyond the idea of Russian or Chinese or, now, perhaps even Iranian hackers gaining access to local or statewide results, there’s the simple question of reliability.

Keene City Clerk Patty Little notes in the case of an actual attempt to thwart the democratic process, it would probably take more than one number being changed on one machine in Keene. In fact, both the state’s relatively spread-out population and its retro election methods are credited as helping safeguard our election from nefarious intent. But that’s not the only danger.

The machines used in Keene, for example, date back to 1993. And while they’re checked before every election, that’s no guarantee they won’t malfunction under the increased stress of counting thousands of ballots on Election Day. We’re not saying it’s likely, just possible. And as past results have shown, it sometimes doesn’t take much to tilt an election one way or the other.

Suppose a single machine in Keene had glitched during the 2016 general election, swapping the results between U.S. Sen. Kelly Ayotte and Gov. Maggie Hassan for Ayotte’s Senate seat. Hassan won each Keene ward by at least 700 votes. Reverse one of those wards and Ayotte would have kept her seat, which was decided by a margin of 1,017 votes statewide.

Now, one might reasonably argue that if just one of Keene’s five heavily Democratic wards had chosen Republican Ayotte by a wide margin, someone would have raised an alarm. But what if, in this example, the same occurred instead in Manchester’s Ward 4 or 5? Ayotte won several wards in that city, and such an error might have gone unnoticed. The makeup of the United States Senate, then, would be altered. One likely result of that single change: The Affordable Care Act would be gone.

Secretary of State Bill Gardner argues local moderators and clerks should not have access to ballots to hand audit machine tallies because they, themselves, are elected in those precincts, and thus could have reason to use the opportunity to their own benefit. Since he is chosen by the Legislature rather than directly by voters, that’s not an issue for him.

We’d counter that voters in a given community choose their local clerks and moderators specifically because they trust them to fairly run the election process; and those officials do handle ballots when transferring them from the machines to Gardner’s office. Further, it seems Gardner’s worry is easily dealt with if any local hand audits are, by law or policy, NOT done specifically on the clerk/moderator’s race. The idea, after all, is simply to verify the ballots show what the machine says.

Little also noted that when hand counts are done during recounts, they’re more likely to be wrong than the machines themselves. That’s a fair point, but if such errors can be rectified during requested recounts — as we assume they are — the same could be done with audits.

We understand there are potential issues with allowing recounts. However, we don’t get why those issues, which we think are easily resolved, are seen to be a bigger deal than the potential of having a machine discrepancy go uncorrected.

One possible middle ground might be to follow Connecticut’s example. The Nutmeg State mandates recounts for any race where the difference falls within a margin of 0.5 percent. Such a policy would increase the number of recounts done, and perhaps help placate those worried that the Granite State’s election machines aren’t being adequately verified.