According to an ongoing Honda audit, Takata engineers have been found responsible for routinely manipulating the results of airbag inflator tests in order to artificially reduce variability.

The news comes courtesy of Brian O’Neill, a former IIHS president, jointly hired by Takata and Honda back in late October to begin the audit.

“We have found examples of what I would call ‘selective editing,’ where they have left out results not because they were bad results, but because the results that remained were better,” said O’Neill. “We found evidence that the report that went to Honda was a shorter version of the original version, and it was a prettier shortened version.”

Findings of data manipulation and misrepresentations could definitely complicate Takata’s efforts to not just restructure, but also secure potential buyers following what turned out to be the biggest safety crisis ever in the automotive industry.

As reported by Autonews, the result of the audit will factor into Honda’s investigation on whether or not they should recall additional Takata inflators, said Chris Martin, a spokesman for the automaker.

“Takata has previously acknowledged and deeply regrets issues related to the integrity of Takata’s inflator validation testing and reporting of test results to its customers,” said company spokesman Jared Levy. “These issues are totally incompatible with Takata’s engineering standards and protocols and entirely inexcusable.”

To make things worse, before Takata agreed to settle on a lawsuit regarding an accident from 2001 when a woman was seriously injured in her Honda Civic, several of the company’s current and former engineers said in depositions that Takata had also altered and misrepresented test data in reports sent to Toyota, Nissan and General Motors.

Since O’Neill’s audit conducted on behalf of Honda and Takata initially focused on airbag inflators validated in North America, Honda has yet to be provided with the final results of this phase, which is said to be nearing completion. Next up for O’Neill and his team is to review as much data as possible, which may take up to two or three months.

PHOTO GALLERY