A major new report by National Academies of Science concludes that there is not enough empirical data to determine whether gun control enhances public safety, or whether gun ownership deters crime. The report calls for further data-gathering on firearms injuries. We suggest that gathering a type of related data is equally critical: how often 911 calls result in the interruption of a crime.
The issue is central to the gun debate. The anti-gun lobbies, while sometimes conceding that people can be allowed to have sporting guns, vehemently opposes gun ownership for personal protection. The lobbies insist that crime victims should rely on 911 instead. For a disarmed victim, the police response to 911 can literally be a matter of life or death. If the data show that 911 won't save your life when you're attacked by a criminal, then it would be difficult for government to claim the moral authority to disarm victims.
We searched for information on the percentage of times a crime-in-progress is interrupted following a call to 911. And we searched for information about how often citizens are protected from harm by police intervention.
There are all kinds of information available regarding 911 calls: numbers of 911 calls made, number of arrests made as a result of calling 911, and types of crimes called in. There are lots of data about 911 response times. For example, Priority One responses in Atlanta and nearby counties take an average of 9-15 minutes. In Washington, D.C., in 2003, the average police response time for highest-priority emergency calls was 8 minutes and 25 seconds. ("Ramsey defends 911 response," Wash. Times, May 11, 2004.)
There are precise data on events such as the two-hour shutdown of 911 in three of New York City's five boroughs on the evening of March 26, 2004 because of phone company problems. There are even data on how many 911 callers are put on hold; the New York Times reported that in Nassau County in 2003, eleven percent of 911 callers got a pre-recorded message and soothing music, rather than a human operator. ("Nassau 911 Callers Are Being Put on Hold," N.Y. Times, Sept. 14, 2003.) In contrast, 911 callers in Quebec City were redirected to an answering machine only about 0.2 percent of the time during a five-month period in 2003. ("Thank you for calling 911, please leave a message," The Record (Kitchener-Waterloo, Ontario, October 22, 2003.)
So why are there no data on crime interruptions?
We looked through the vast wealth of criminological information at the U.S. Department of Justice website, and we looked through print-based resources. Not finding any statistics anywhere on violent crime interruption by the police, we asked the statisticians at the Department of Justice directly.
One day later, we received the following answer from the DOJ's Bureau of Justice Statistics: "I'm sorry but the Bureau of Justice Statistics (BJS) does not collect data on law enforcement intervening or preventing crimes that are in progress."
The Canadian government does not gather such statistics either, even though the Canadian government vehemently insists that citizens must not use firearms to protect themselves or others. (The non-existence of the Canadian data was confirmed for us by M.P. Garry Breitkreuz, Deputy House Leader for the Official Opposition in the Canadian Parliament, based on his queries to the Library of Parliament and to Statistics Canada.)
Although we were unable to find the statistics for interrupted crimes, we did find a study of how many criminals are caught after perpetration of the crime. However, the most recent research is more than two decades old.
In 1977, the Kansas City (Missouri) Police Department examined variables affecting police response time to 911 calls. The study concluded that the factor which most hampered the effectiveness of the 911 system was not police response time, but citizen delay in alerting the system.
William Spelman (a Professor at the University of Texas's LBJ School of Public Affairs) and Dale K. Brown showed that the Kansas City results could be replicated in other cities. In their 1981 book Calling the Police: Citizen Reporting of Serious Crime, Spelman and Brown selected four additional other cities to study, each having significant regional, policing, and population differences: Jacksonville, Peoria, Rochester, and San Diego. Despite the differences, the outcome measures were almost identical among all four cities studied.
Spelman and Brown confirmed the Kansas City results: the most important reason criminals escape, despite a call being made to 911, is that the call is made too late. In other words, the police were exonerated. The police were not, in general, failing to respond quickly to 911 calls; the calls simply came too late to do any good. (Of course there are horror stories of negligent and torpid police response, but these disasters represent the exception, not the rule.)
The Spelman and Brown report had important implications for the allocation of police resources: putting more money into speeding up police response times to 911 would be too expensive and would offer insufficient benefit to justify the expense. As Spelman and Brown found, "arrests that could be attributed to fast police response were made in only 2.9 percent of reported serious crimes."
According to Spelman and Brown, if the crime was reported while still in progress, the arrest rate was 35 percent. If the crime was not reported while in progress, and the victim took 60 seconds to get to a phone, the arrest rate dropped to 10 percent.
Now of course making an arrest is not the same as stopping a crime in progress. If the police are called while a murder is taking place, they may (about 35 percent of the time) arrive in time to arrest the murderer, but not necessarily in time to save the victim's life.
Yet even if we made the artificial assumption that every arrest meant that the crime in progress was thwarted, we see that two-thirds of the time, the police will not arrive in time to protect you.
Nevertheless, the gun prohibition lobby, the District of Columbia government, and many government officials, insist that victims should not protect themselves with firearms. They must instead rely on 911.
That command ignores the fact that any criminal in control of a crime scene will not permit his victim to call the police, and that the neighbors may be unaware of the crime in progress.
Moreover, even if the police are alerted immediately, they still have to spend time traveling to the scene of the crime, although the victim may need help within seconds.
For example, on June 5, 2002, eighty-nine year-old Lois Joyner Cannady called the Durham County, North Carolina, 911 to ask for immediate police aid. She was killed before the police arrived on the scene. Police deputies came within minutes, but the killer was long gone.
Might the outcome have been different if Mrs. Cannady had reached for a gun instead of a phone? Two 80-year old women homeowners did just that, in Elbert County, Georgia. A News Channel 32 report stated that, according to Sheriff Barry Haston, "having the guns kept those women alive." Haston said, "In these two cases I'm actually glad they did because it could have been a different story if they didn't." There are many other reported cases of persons as old as Mrs. Cannady, or older, using firearms successfully for protection.
When potential crime victims (i.e., everyone) consider whether to adopt particular defensive measures (locks, guns, window bars, alarms, etc.), they must make trade-offs of costs and benefits. For example, window bars might prevent a criminal from coming in, but they can also block the exit in case of a fire. For citizens to make well-informed decisions about self-defense, citizens ought to know how likely it is that the government will rescue them in an emergency.
We cannot expect perfection from the police; after all, they travel by automobile or by foot, not by teleportation. We can expect that government or university researchers (many of whom are heavily subsidized by the federal government) would gather statistics directly relevant to life-or-death decisions.