Crimes Uncounted
How Philadelphia police his rape complaints
Sidebar to the second of two parts

After FBI questioned one tactic, another was found

By Mark Fazlollah, Michael Matza and Craig R. McCoy,
INQUIRER STAFF WRITERS

Part 1: Women victimized twice in police game of numbers
Part 2: How police use a new code when sex cases are 'unclear'
Police used 'throwaway categories' since 1960s
Rape squad office? 'It's sort of scary'

Perhaps the most breathtaking example of statistical manipulation by the Special Victims Unit occurred during the first few years of its existence.

Mayor William J. Green established the unit in 1981. Police reported a 10 percent decline in rapes in 1982, followed by a 2 percent drop in 1983.

Rapes had not really declined, however. Police were keeping the number down by declaring large numbers of complaints "unfounded" - groundless.

After the FBI privately raised questions about the practice in 1983, the "unfounded" numbers dropped sharply almost overnight - not because sex-crimes investigators were taking women's complaints more seriously, but because they had found a new hiding place for them.

The public never knew what happened. "Unfounded" data were not published at the time.

The story is laid out in confidential correspondence between the FBI and the Police Department obtained recently under the Freedom of Information Act.

In 1983, the documents show, Philadelphia police deemed rape complaints "unfounded" at a rate far exceeding that of most other departments.

In the first half of that year, 52 percent of all rape complaints were dismissed as fabrications. The national average was about 10 percent.

Read the FBI letter

In December 1983, a top FBI official requested an explanation of Philadelphia police procedures, "since they apparently deviate from those of similar agencies." The bureau also pointed out that it would soon, for the first time, begin publishing yearly "unfounded" rates for major cities. The message: Philadelphia had better clean up its act.

The sex-crimes unit defended itself, saying that no legitimate complaints had been rejected.

Read the city police response

In a January 1984 letter to the FBI, Chief Inspector Thomas Roselli, who oversaw the sex-crimes unit, said the high "unfounded" rate reflected a new policy of accepting all complaints, regardless of how "vague or ill-defined" or even "wildly improbable" they were.

With so many dubious complaints in the system, Roselli suggested, it was inevitable that the number of "unfoundeds" would rise.

Roselli devoted most of his four-page, single-spaced letter to scenarios in which women lied about rape. He listed two dozen of them.

Here are a few:

"Complainant who suffers from a medically diagnosed mental condition reports a rape . . . penetration by extraterrestrials, evil spirits, television or movie star."

"Complainant reports rape to cover for infidelity, indiscretion, lateness, pregnancy, etc."

"Complainant reports rape in order to make husband/boyfriend feel guilty or jealous. Usually reported following a 'lover's quarrel.' "

While defending its conduct, the unit did take steps to slash the unfounded rate. It accepted more complaints as rapes, driving the rape total up 14 percent in 1984.

At the same time, investigators began threatening complainants with lie-detector tests so that women suspected of fabricating their stories would withdraw complaints. Police even arrested some women for filing false complaints.

News of these tactics "spread like wildfire," Roselli, now 72 and retired, said in a recent interview. "Word got out that we were arresting people."

Simultaneously, the unit developed another strategem. It began putting hundreds of sexual-assault complaints in a bureaucratic twilight zone - Code 2701, "investigation of person" - to keep them out of the city's crime statistics.

The result was a tour de force of statistical manipulation.

In 1983, investigators "unfounded" 635 of the 1,464 rape complaints they received.

In 1984, the number of "unfoundeds" dropped 71 percent - to 183 out of a total of 1,128 complaints.

Had all those cases simply vanished?

No. They had been been moved somewhere else - to "investigation of person" - where they did not show up either as rapes or as "unfoundeds."

In some subsequent years, the rape squad's unfounded rate - once among the highest in the nation - was among the lowest. The problem seemed to have been solved.

But beneath the numbers, nothing really had changed.