​

MEASURING the effectiveness of surveillance cameras in crime prevention is challenging. A drop in crime may be influenced by citywide trends or factors other than the cameras. It is difficult to control for those other variables and to find a similar area that can serve as a control group.

Despite these challenges, there have been several past attempts to assess surveillance system effectiveness, notably in the United Kingdom. More recently, two groups have conducted studies of these systems in the United States.

The first of the new studies comes from the University of California, Berkeley’s Center for Information Technology Research in the Interest of Society (CITRIS). It examined San Francisco’s Community Safety Camera program, and fulfilled a requirement of the 2006 codification of the program that was established a year earlier. The program placed 71 unmonitored cameras in high-crime areas in the city.

CITRIS had access to the data and stakeholders, which is key. “A major barrier to any kind of research effort to examine camera systems is a reluctance to bring [outsiders] in to do an assessment and a lack of funds to support it,” says Deirdre K. Mulligan, one of the study’s authors.

“It really is a feather in the city’s cap that it wanted not to just deploy a technology,” says Mulligan, but “to build into the system an assessment.”

In addition to the access and the choice to do both a qualitative and quantitative analysis, a strength of the study was the staggered implementation of cameras, which allowed for “quasi” control groups. This enabled the researchers to use changes in crime rates in areas that did not receive the cameras until later as a basis for before-and-after comparison.

CITRIS found that “Part 1” felonies committed in public declined by 30 percent from the time cameras were installed in 2005 through January of 2008. The most notable decrease was in larcenies (such as pickpocketing and purse snatching) in surveilled areas.

Violent crimes did not decline, although that was one of the stated goals of the cameras.

The project’s other main goal was to provide the San Francisco Police Department (SFPD) with a forensic tool. The cameras proved useful in helping the SFPD solve crimes. Officers requested footage about three times per month; the footage assisted SFPD in charging a suspect in six cases over the three years examined.

Another recent study looked at the addition of cameras to a private housing complex, Peter Cooper Village, in Manhattan. It was selected in part because its population was significant and an adjacent complex (Stuyvesant Town) could serve as a control group, according to the authors, Jeffrey B. Roush, an operations management consultant, and New York University sociology professor David F. Greenberg.

Roush, previously had worked at the property, which enabled “access to the crime counts, and we verified that the property’s internal crime database had enough integrity to perform this type of intervention study.”

Not all properties have that type of solid internal reporting data, and even if they do, they may not be willing to share it with researchers, explains Roush.

The study covered three years before cameras went in and three years after. However, a fundamental limitation was that the property had a low overall crime rate even before the installation of the security cameras. Thus, it may not be surprising that the surveillance system did not yield dramatic results.

This study found only a fractional decline in larcenies. The authors acknowledge that the drop could not be said to show definitively that the project had an impact on crime.

More interesting was that the authors used a methodology called the Poisson regression (rather than some of the more traditional CCTV study techniques), which Greenberg says is more accurate where crime rates are low.

Roush says that the findings illustrate how this methodology could be used to assess effectiveness of camera projects. He says either government or third-party organizations should devise recommendations on how to conduct such research. Roush adds that a framework for making an assessment must be set out before the project is implemented.

Often, Roush says, citizens are left saying, “You spent millions of dollars on this and how effective is it?” In the end, “everyone just kind of shrugs and puts their hands in the air, [saying] ‘well I don’t know.’”