Rarely to get exciting reading articles in academic journals (whether that says bad thing about me or the journals I leave to you), but this is exciting: "The Philadelphia Foot Patrol Experiment: A Randomized Controlled Trial Of Police Patrol Effectiveness In Violent Crime Hotspots." It's in the current issue of Criminology (like most academic journals, unavailable to the general public).
The Newark Foot Patrol Experiment was researched in the 1970s. And though it showed foot patrol in a more positive light than many people remember it for, it was hardly the unequivocal support for foot patrol I would have expected.
Now I know foot patrol works, but get a bunch of academics in a room and ask a simple question like, "Does foot patrol work?" and you'll get a lot of "we don't know" and "no" and "more research is needed." Even in the police world, opinion is split.
Well finally somebody has done a proper study of foot patrol. The bottom line? In high-crime areas, foot patrol decreased serious violent crime by 23 percent. This happened just after three months of foot patrol. No big difference was found in lower-crime areas, but then we fall back on the Newark Experiment and reduced public fear.
I bet you'll never hear of a study showing the crime-reduction benefits of officers remaining "in service" to answer radio calls.