With another round of 10-Ks and 10-Qs mostly filed through the early part of 2015, experts are assessing the quality of XBRL filings and seeing not much improvement.

Scale errors continue to be the most common problem found in XBRL filings, says Alex Rapp, co-founder of Calcbench, an interactive data platform that leverages XBRL-formatted financial statement data. A scale error occurs when a company means to report a number as 325 million, for example, but instead records it missing a trio of zeros, so it appears as 325,000. “Unfortunately, many filers don’t appear to realize this is an issue, and therefore are not checking for it,” wrote Rapp in a recent informal assessment of XBRL errors. “This should not be taken lightly.”

The Securities and Exchange Commission has taken heat on multiple fronts, including from Congress, for not getting more proactive in compelling companies to clean up error-prone XBRL filings. Companies are not required to have their XBRL filing audited, and no outcry from the marketplace has suggested it should be required. The House Financial Services Committee is considering a bill to exempt companies with revenue below $250 million from the XBRL filing requirement. Groups like the U.S. Chamber of Commerce are calling on Congress to come up with a better method of creating interactive, searchable market data.

Calcbench has been monitoring errors and correcting them as much as possible as the data populates its interactive platform, including incorrect dates, incorrect numbers, and other quality issues. Scale errors appear most often in share counts, noted Rapp, and they are common in the text of footnotes to financial statements. “But these errors can happen anywhere,” he wrote. “We get one or two filings every period where every value is in the wrong scale.”

Scale errors have abated only a “small amount,” according to Rapp, with around 10 percent of all filings containing such mistakes. “Most every company has done this at one time or other, and we see it happen across all filing software suites and service providers,” he wrote. The best way to check for scale errors, said Rapp, is to look at a raw list of numbers in the filing after they are prepared. “All filing software suites are different, and their rendering will oftentimes be deceiving.”

The document and entity section of a filing is another common trouble spot, said Rapp, where companies provide basic information about the entity. Error rates in this section have been cut about in half, he said, declining from 5 percent of filings the past few years to 2 percent more recently. Finally, companies routinely give the wrong date for their fiscal year end, giving a date instead that matches the quarter end rather than the year end, or associate the date with the entity shares outstanding. “Companies often times forget to update this value from the previous filing,” Rapp wrote.