Universal background check legislation stumbled forward in the U.S. Senate yesterday and collapsed in Washington state, despite still-strong poll numbers on the issue. But the conventional wisdom on just how many American gun sales go unchecked might be based on a myth of bad data.

In its first vote on the issue since it became such a political priority for the administration, the Senate Judiciary Committee yesterday advanced a proposal from New York Senator Chuck Schumer that would require a background check for any gun sale, regardless of venue. Currently, only sales in gun stores require such checks, though they may not be as lacking as typically assumed. The Schumer bill came before the committee after an effort at bipartisan compromise collapsed last week. As CNN notes, the committee vote was 10-8, along party lines — not a positive sign for its being passed by the full Senate.

The news for proponents of background checks was even worse in Washington state, where the state House all but ended an effort to enact universal background checks. Washington's state legislature is strongly partisan, with Democrats holding a large majority.

Both votes happened on the same day polls released by Pew Research and the Washington Post/ABC indicated that the measure still sees broad support among Americans. The Post/ABC poll suggested that over 90 percent of respondents favor universal background checks; Pew put that number at 83 percent — but also found that 74 percent of NRA members support the effort.

(Pew also created an excellent interactive graph of its gun data that's worth a look.)

What isn't clear is how many transactions an expanded background check system would cover. Because sales at gun shows and between two parties are unregulated, it's not clear exactly how many occur. The Associated Press looked at one commonly used statistic suggesting that as many as 40 percent of sales happen without regulation and found reason for skepticism.

That figure, the AP notes, is rooted in a 1994 poll of 2,500 respondents, of whom 251 had purchased a gun. 36 percent of those sales occurred outside of traditional gun stores. That's where the 40 percent figures comes from:

But the study's researchers found considerable ambiguity and some apparent contradictions in the responses. With a clear picture eluding them, they estimated 30 percent to 40 percent of the acquisitions were off the books and would not have been subjected to a background check.

Only 4 percent of gun sales were thought to have come through gun shows or flea markets — a corner of the market that is a top concern today for those who want to expand background checks to close the "gun-show loophole," as Obama's proposals would do.

Furthermore, 17 percent of the sales, that study found, were between family members.

Bad data on gun transactions is endemic, in part because the NRA has consistently worked to impair the government's ability to collect it. That bad data works both ways. It makes it hard for advocates of increased measures to cite accurate statistics — and it makes it hard for opponents to do the same. This leaves legislators to largely make decisions on instinct and under pressure from advocacy organizations.

If those poll numbers are right, there's only group that seems unable to get any political traction: the voters.