You don't need Tor. You don't need super-secret encryption. And you definitely don't need a background check. Getting firearms online, fast, without any legal fuss, is pretty easy. In fact, a new report concludes plenty of domestic abusers and violent felons are getting guns that way.

Titled "Online and Off the Record," the report by Everytown—the Mike Bloomberg-affiliated pro-regulation gun group—suggests that hundreds of thousands, probably millions, of guns nationwide are now being sold online sans paperwork or background checks, and as many as 10 percent of them are going straight to convicted criminals who are legally barred from owning firearms.

The report focused on web-based gun sales in Washington state, where voters will decide in November whether to require background checks on all weapons purchases and where "more than forty thousand guns are posted for sale by unlicensed sellers on just five websites" each year, without a background check requirement, Everytown estimates.

Here are the five biggest sites buyers are using:

The most jarring stat in the report—it's also likely to be the most disputed—is Everytown's estimate that nearly 10 percent of those untracked online purchases went to felons or others who can't legally own the guns. Investigators focused on "guns wanted" ads from prospective buyers and focused on 81 that included identifying info. Through court records and personal interviews, Everytown found that 8 of the 81, or about 10 percent, "had been convicted of crimes that prohibited them from possessing firearms," particularly domestic abuse.

That may not seem like a large-enough or representative sample.* Still, finding eight convicts on a gun-selling website that easily is pretty alarming.

The report details the case of one man, from King County, who lost his gun rights after a felony car theft conviction in 1976. About a decade later, he was hit with a restraining order after an ex-wife said he'd threatened "to go buy a gun so he could knock me off." The following year, he pled guilt to domestic violence assault; in 1992, he allegedly assaulted two cops. And this May, he was on Armslist.com, trying to buy a Browning 9mm Hi-Power: "Cash in hand for the right deal," he wrote.

Not all of the online gun-sales ads are created equal. Many are for lever action Marlin rifles or cheap .30 caliber single-shot bolt-action long guns—types that are likelier to be used for target-plinking or varmint-hunting than mass-killing. But there are also 9mm Beretta handguns. Glocks with laser sights. Snub-nose revolvers, AK-47 reproductions, and Mini-14 rifles with tons of ammo and magazines. If someone wanted a quick killing weapon, no questions asked, this would seem like a good way to go about it.

This election day, Washington voters will make a decision on Amendment I-594, which would require across-the-board background checks. At a poll in April, voters were overwhelmingly in favor of the measure, 72 percent to 19 percent. The only problem is that a majority—55 percent—also favored I-591, a measure that would bar any new background checks.

Should the two contradictory amendments both pass, it's unclear which one, if any, would actually become state law. But online gun sales would likely spike while the legalities get decided.

Update: An Everytown researcher who read this post emailed to offer some context on their numbers methodology:

In addressing what are going to be inevitable questions about the sample size, you can refer people to footnote 21 in the report:

Based on the sample size of 81 examined buyers, the margin of error is 3.4 percent to 16.4 percent.

The confidence interval is calculated as P +/- 1.96*[(P*(1-P)/n) ^.5 = 9.9% +/- 1.96 (.099*.901/81)^.5 = 3.4% to 16.4%.

In English, this means that there is no way the sample we achieved could have occurred by chance alone. And even at the lowest bounds of the confidence interval, the share of criminal buyers is dramatically (5x) higher than the share of criminals trying to buy guns at Washington dealers — and online in unlicensed sales, there is no background check to stop them.

There's already a lively discussion in the comments about the methodology, so I'll leave this here for interested parties to parse as they see fit.

[Images via]