In the latest in a series of audits questioning the accuracy of New Orleans crime statistics, Inspector General Ed Quatrevaux’s office issued a report Wednesday that says the New Orleans Police Department misclassified nearly 40 percent of a sampling of robbery cases the office reviewed.

As with the earlier audits, which dealt with rapes and thefts, the report drew a spirited rebuttal from Ronal Serpas, who retired suddenly as NOPD superintendent last week but was still in the department’s top spot when the audit was completed. Serpas suggested the report contained major errors and was broadly misleading.

Regardless of whether the audit is accurate, it does leave open perhaps the most important question when it comes to robberies: If the NOPD is misclassifying them, what effect is it having on the crime rates the city reports to the FBI each year in the so-called Uniform Crime Reports?

Records show that the NOPD reported 1,065 robberies in 2012. But the IG’s report — which surveys a three-year period ending May 31, 2013 — focuses on a very small slice of the roughly 3,000 robberies that the department reported over that time, giving no definitive answer to whether classification problems are keeping New Orleans’ crime rate artificially low.

The IG examined 30 cases in which police filled out a “change of signal” form — that is, cases in which officers determined at least 24 hours after a robbery was reported to police that the crime should be counted as something other than a robbery.

Of the 30 cases the IG scrutinized, the audit found that 11, or 37 percent, were wrongly reclassified as “miscellaneous incidents” rather than robberies. Such incidents do not get counted as crimes by the FBI.

But a very small fraction of the cases the NOPD deals with involve such change-of-signal forms, a fact not addressed in the IG’s report or in the city’s response. City officials could not provide an exact number or percentage Tuesday of how many calls for service wind up resulting in such a form being filed.

Serpas’ response to the report suggested that, even if one agreed with its findings — which he did not — the audit exposed an error rate of only about 0.1 percent in NOPD’s reported robbery statistics. The response does not explain how Serpas arrived at that number, but the former chief added that the FBI considers an error rate of 10 percent or less acceptable.

In addition to cases involving signal changes, the IG’s report also looked at 45 cases in which citizens called the department to report robberies that police ultimately determined were “unfounded,” deciding that no crime had actually occurred.

According to FBI rules, departments are supposed to count such calls in tallies of robberies reported to police, but then to list them separately as “unfounded.” The NOPD did not do that.

The distinction likely doesn’t matter much to the public because the final tally of robberies counted by the FBI — called “number of actual offenses” — does not include the unfounded ones.

In other words, the NOPD may not have filled out its paperwork correctly, but the error in counting “unfounded” crimes would not affect the city’s reported robbery rate or its overall crime rate.

The department is now reporting such incidents, often involving tripped burglar alarms and the like, in its monthly reports to the FBI.

Perhaps the most troubling finding in the report had to do not with the recording of statistics but the maintenance of evidence, a long-standing problem in the department. The report says that 26 of the robbery cases reviewed made mention of physical evidence. But in four of those cases — or 15 percent — the evidence was not submitted to the NOPD’s Central Evidence and Property Section.

The department did not dispute that finding.

Serpas’ six-page response to the report is pointed at times and accuses the inspector general of making a host of errors. For instance, while the IG complained that records from 2010 and 2011 were not available, Serpas’ letter claims the department offered the records, but only on paper, and “instead of doing the extra work required to analyze our paper records,” the IG simply “acted as if the data did not exist.”

Much of Serpas’ response is devoted to quibbling with the IG’s findings in specific cases, but he also notes that he created a special unit in 2010 tasked with ensuring the accuracy of crime statistics. That unit, he writes, had already flagged and corrected many of the robbery cases the IG said had gone unreported.

“Given research conducted in other jurisdictions on the complex and sometimes controversial world of UCR crime reporting, there is reason to believe that NOPD reports UCR crimes more accurately than many other police departments throughout the country,” he wrote.

Quatrevaux remains skeptical. The new report is the third released by his office that raises similar questions, with earlier reports focusing on rape statistics and thefts in the French Quarter. The state’s legislative auditor also issued a report last year raising questions about the NOPD’s crime reporting, and newspaper stories have taken aim at the department’s accuracy.

The report released Wednesday notes that New Orleans’ robbery rate is suspiciously low — “46 percent lower than 24 other cities with the highest crime rates” — and Quatrevaux said in a phone interview that he saw a similar pattern with rapes.

“From 1995 to 1999, the number of rapes here was larger than the number of homicides reported, and that’s characteristic of the whole country,” he said. “But beginning in 2000 or 2001, suddenly the number of rapes dropped below homicides, and it has stayed that way. And that’s because we’re not counting them correctly.”

Quatrevaux acknowledged that the latest audit doesn’t answer how dramatically the misclassifications it identifies are skewing the crime rate. To figure that out would require an extremely broad audit, he said.

“But our stats are demonstrably different from other cities, and there needs to be an explanation,” he said. “What’s important here is getting numbers we can compare to other cities, numbers that follow the FBI’s guidance. Until then, we’re just talking.

“Going forward, though, we’ve got our eye on this. As soon as I can talk to the new chief, I’m going to point out the importance and solicit his cooperation in fixing this problem.”

Follow Gordon Russell on Twitter @gordonrussell1.