Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Surveillance Cameras topic
No spam. Unsubscribe anytime.
UC Berkeley preliminary study finds property-crime declines near San Francisco surveillance cameras
Summary
A UC Berkeley/CITRIS preliminary analysis presented to the Police Commission found a measurable drop in property crimes within about 100 feet of city pilot cameras, but researchers and commissioners urged caution: the effect is preliminary, concentrated on property offenses (especially thefts from vehicles), and the report showed little consistent effect on violent crime.
Steven Raphael, a researcher from the University of California, Berkeley's CITRIS lab, told the San Francisco Police Commission that the team's preliminary statistical analysis finds a statistically measurable decline in property crimes within roughly 100 feet of 19 pilot camera sites. "We're finding a relative decline of 22% in the area near the cameras," Raphael said during a presentation to the commission.
Raphael described the study design and data: the researchers used incident data provided by SFPD covering roughly 59,000 incidents from Jan. 1, 2005, through Jan. 28, 2008, across 19 camera sites. The team compared changes in average daily crime rates inside camera view to changes at immediately adjacent areas (used as comparison samples). The largest and most statistically robust declines were in property crimes within the camera'view; among property subcategories, larceny-theft showed the clearest decline, driven largely by thefts from vehicles (about 51% of the larceny category in the dataset).
Raphael cautioned the commission that the results are preliminary and sensitive to methodology choices. He recommended additional data (longer pre- and post-installation windows, arrest data tied to camera-derived leads, and qualitative interviews) and warned that pooled averages can hide site-by-site variation. "These are preliminary findings," he said. "We're going to extend the data back a little bit further and forward a little bit so we can have an entire calendar year on either side of a camera installation to control for seasonal patterns."
Commissioners, city staff and community members pressed for more detail on several fronts: whether arrests that derived from camera footage could be added as an outcome, whether differences in camera frame rate or visibility bias results, and whether cameras simply displace some violent incidents to nearby areas. Some commissioners said the early pattern did not justify expanding the pilot immediately; others said the program remains worth evaluating further. Several commissioners said an existing moratorium on new camera installations should remain until the final report is complete.
The presentation and Q&A took place as part of agenda item 5; the researchers emphasized that final findings will include additional empirical tests, cluster-level estimates and qualitative follow-up to better explain homicide and violent-crime patterns that appeared in the preliminary numbers.
