An extraordinary feature of the 2012 elections has been the barrage of outside money unleashed on America’s airwaves. Deep-pocketed groups such as super PACs and nonprofits rival the campaigns themselves as sources of TV advertising, and reporters are scrambling to shine some light on where all the cash is coming from and how it is being spent.

To measure the role of money in the TV ad wars, the media have so far relied primarily on two sources. One is the Federal Election Commission, which requires these organizations to make monthly or quarterly reports on their expenditures; the other is Kantar Media, a private research group that scours the airwaves for TV ads and estimates their overall costs, then sells that information to political clients and major media outlets.

Here is the interesting part: the two sources’ numbers don’t add up.

Not that they should match dollar for dollar; Kantar and the FEC have different missions and methodologies. But the differences in their numbers aren’t minor. They are discrepancies in the tens of millions of dollars.

For example: CJR’s breakdown of FEC data found twice as much TV ad spending by two leading super PACs, Priorities USA (pro-Obama) and Restore our Future (pro-Romney), than Kantar found—as you see in the chart below. (This is as of September 20—we stopped before the latest FEC disclosures so both sources would cover approximately the same time period). In other cases, the divergence is in the opposite direction: Kantar showed the conservative nonprofit group Americans for Prosperity spending more than $7 million more on TV ads than the FEC showed.

Sources: Federal Elections Commission, data accessed via the Sunlight Foundation on Sept. 20, 2012. Broken down by “Purpose” to isolate TV ad spending. Kantar Media, accessed via The Washington Post’s Mad Money campaign ad tracker on Sept. 20, 2012.

On one level, this is a cautionary tale for reporters, who often treat information from the two sources as if they were hard numbers from similar places. In fact they are soft numbers from very different sources, and campaign-spending stories should reflect that reality.

On a deeper level the massive variations between the numbers expose a troubling situation: Less than two months away from electing a president, we don’t really have a handle on how much outside groups are spending on television campaign ads to influence that choice.

“What we want in disclosure is hard numbers that add up and match,” said Bill Allison, the editorial director of the Sunlight Foundation, a nonprofit that promotes transparency in government. “With TV advertising we just don’t have that.”

There are a number of perfectly good possible reasons for the very different numbers Kantar and the FEC often provide. Among them: Kantar measures ads that have aired and not future buys, while the FEC counts spending toward future ads; Kantar does not include local cable stations, while the FEC does. And except for a short time window before the election, the FEC does not require nonprofit groups to disclose so-called “issue ads.” Then there is the possibility of underreporting, or mis-reporting to the FEC. But as to which reason applies in any given case, who knows? “What’s really scary is I’m basically guessing,” said Sunlight’s Allison.

There is more below on the differences in methodologies between the two entities. But essentially, Kantar monitors a large but incomplete subset of America’s airwaves for campaign ads as they run, while the FEC monitors a large but incomplete subset of groups’ outlays as they spend.

The bottom line is that uncertainty about ad spending allows super PACs and nonprofits to continue operating in a legal and financial fog, which hinders the press in seeking to hold them accountable.

CJR calculated the disparities by comparing Kantar statistics used in The Washington Post Mad Money campaign ad tracker with FEC expenditure data compiled by the Sunlight Foundation. We broke down the FEC totals by media category to isolate TV ad spending.

The divergence between the media’s two crucial sources on TV ad spending has sometimes resulted in coverage that misleads readers. News organizations often cite Kantar as a measure of overall TV ad spending without noting the gaps in its data, and occasionally refer to Kantar and FEC data interchangeably, without explaining that their figures vary widely.

Sasha Chavkin covers political money and influence for CJR's United States Project, our politics and policy desk. He has written for ProPublica, the Center for Public Integrity, and The New York World. Follow him on Twitter @sashachavkin.