LAPD’s Response Time Now Reported as ‘Uniformly Long’

The principal author of a manpower study for the Los Angeles Police Department, declaring that his work has been misinterpreted, said this week that police response to emergency calls is “uniformly long” citywide--not skewed in favor of affluent areas, as previously reported.

He also said he found that police typically took 9.2 minutes to arrive at an emergency anywhere in the city, not 12.3 minutes, which was widely reported as one of his key findings when the the study was released last week.

Peter Bellmio, manager of criminal justice projects for Public Administration Service, a private, nonprofit consulting firm in Virginia that was hired by the Police Commission, made the statements in a letter to the Police Department, which the department made public Friday.

Bellmio’s letter clarified data being used in a longstanding controversy--which the study helped ignite anew--that the LAPD responds quicker to emergencies in the largely Anglo San Fernando Valley than it does in the largely minority South-Central portion of the city.

The report actually concluded that the Police Department takes too long to respond to emergencies but does not respond appreciably quicker in any of the city’s broad geographic areas. It said emergency response time should be no more than 7 minutes.

Bellmio declined to be interviewed about his report, saying he was contractually bound to make no public statements before a March 8 meeting of the Police Commission, at which his report will be considered.

He said in his letter to the Police Department that confusion arose over a 12.3-minute emergency response figure that was highlighted in a chart within the nearly 200-page report as an “average.”

For technical reasons, that “average” figure was, in fact, misleading, Bellmio said.

“Average travel times did not accurately measure the speed at which emergency calls are being answered,” he said.

Times Skewed

Researchers found that average times were badly skewed because police officers sometimes forget to notify headquarters promptly when they arrive at the scene of an emergency.

Police officers dispatched to emergencies can notify headquarters by radio or, more commonly, by pushing a button, marked “on scene,” on a small computer console in their cars, officers said.

However, the Public Administration Service found in its report that “in some cases, officers rightly pay more attention to events at the scene of emergency calls than they do to logging changes in their status. Consequently, the time at which units arrive at the location of a call for service may not always be recorded accurately.”

Emergency response time is made up of two factors--the time dispatchers take to field a call and assign a patrol unit, which is known as “dispatch delay,” and the time it takes for a patrol unit to get there, which is known as “travel time.”

Researchers indicated in the report that they discovered the shortcoming with “average” figures for response times when they attempted to find an average for the travel-time factor alone.

They came up with “unusually slow and long travel times,” the report said--a problem “created by inaccurate arrival times entered in the field by officers.”

To come up with what they said was a more realistic travel time en route to emergencies, which the researchers needed make deployment calculations, they said they decided to use a different statistical designation--a median, as opposed to an average.

The median is the middle value of a group of numbers arranged in order of size.

Thus, “the median time represents the value at which 50% of the calls have a slower or faster travel time,” Bellmio said in his letter.

By relying on median rather than average travel times, the researchers were able to minimize the warping effect of those individual travel times that they apparently believed were absurdly long.

“Median travel times more accurately describe speed in responding to emergency calls,” Bellmio said in his letter.

Included in the letter was a new chart, which calculated the median travel times and dispatch delays for police emergency responses in each of the city’s 18 areas.

(The report itself included a chart that showed these median travel times, but not in combination with the dispatch delays. That chart was thus not useful for calculating overall response times and was widely overlooked.)

While the new chart states that median response times vary by as much as 2.7 minutes from one area to another, it says they are about the same overall in each of the city’s four broad geographic regions.

Asked to comment on the letter, the Police Department’s chief spokesman, Cmdr. William Booth, said:

“We’re pleased that the Public Administration Service took the time to send the letter to caution people that it (the report) was probably being overly interpreted, and not always accurately, according to what they had been reading about or hearing about it in the news.”

The letter reiterated the study’s principal conclusion, that the cause of long response times is dispatch delay caused by lack of patrol units available to handle emergencies.

The widely quoted chart in the report, which showed average dispatch delays, travel times and overall response times, was used by the authors to illustrate the dispatch delay phenomenon. It was in fact criticized elsewhere in the report as a poor means of depicting typical travel times. However, these later references were overlooked in initial news reports.

The average response times were calculated from July 1, 1986, to June 30, 1987, and the study showed that it took an average of 12.5 minutes to respond to an emergency in the Central Bureau, 12.3 minutes in the West Bureau, 13.3 in the South Bureau and 11.4 in the Valley Bureau.

Median response times were calculated for the same period and the study showed that it typically took 9.2 minutes to respond to an emergency in the Central Bureau, 9.6 minutes in the West, 9.4 in the South and 9.3 in the Valley.

These times “more closely depict reality,” the authors of the study said.

However, after the study was released, Los Angeles Police Chief Daryl F. Gates issued figures of his own, which relied on yet another statistical method to show that, for the month of January, response times were even better.

Gates showed news reporters a computer printout that said it typically took police 7.1 minutes to respond to an emergency in the Central Bureau, 8.7 minutes in the West, 8.1 in the South and 8.3 in the Valley.

Booth said the chief’s figures were arrived at by averaging response times for the middle 80% of all calls--meaning that the fastest 10% of all calls and the slowest 10% were dropped from calculations as unrepresentative.

Gates told reporters that the new figures resulted from unspecified new strategies and recent improvements in dispatching procedures.

Two dispatchers who agreed to be interviewed on condition that their names not be published said there had been no recent dispatch changes.

COMPARING EMERGENCY RESPONSE TIMES

Consultants studying deployment of Los Angeles police suggest that newly released median response times reflect a better analysis than previously released average response times.

Average Median Response Response Time Time (in minutes) (in minutes) Central Bureau 12.5 9.2 Central 11.6 8.3 Rampart 12.1 9.4 Hollenbeck 12.0 8.9 Northeast 13.5 10.4 Newton 13.1 9.6 South Bureau 12.8 9.4 Southwest 13.1 9.4 Harbor 11.0 8.7 77th St. 13.1 9.8 Southeast 13.3 9.1 West Bureau 12.3 9.6 Hollywood 11.7 9.3 Wilshire 13.0 9.8 West L.A. 11.9 10.0 Pacific 12.4 9.3 Valley Bureau 11.4 9.3 Van Nuys 12.1 9.6 West Valley 11.1 9.5 N. Hollywood 9.9 8.0 Foothill 12.6 10.7 Devonshire 11.2 8.5

Source: Public Administration Service