Q:How true are online community crime rankings?

Such was the question one reader posed to The Herald regarding an Internet site that rated Farmville as safer than just 12 percent of cities in the United States.

“They’ve got information … but they don’t have warning labels. They don’t have explanations to what their data says and so they leave it up to the imagination and for individuals to kind of make an assumption,” Farmville Police Chief Curtis Davis said.

The particular site that led to the question, for example, offers a violent crime comparison per 1,000 residents. The chances of becoming a victim, it figures, are one in 475. It also lists

Farmville as having had one murder. The last time Farmville had a murder was in 2009 the year there was a quadruple murder on First Avenue.

To have one murder a year (in all, there were six murders in Farmville technically listed in 2010 stats) and none since would seem to represent a six-year average.

Davis points out that the FBI and State Police compile crime numbers. They are compiled for tracking “but there is a warning label on both of those numbers not to use that data to rank a town or city in regards to safety. So the people who designed the collection of data actually issue a warning not to use the data for the very purpose it’s being used here.”

Data can tell a different than intended story.

An aggressive retailer pursuing shoplifters can lead to more shoplifters being caught, but the figures can also be applied to estimate the likelihood someone could become a victim of a property crime, even though the two seem unrelated.

The police chief noted that if you’re going to “rate crime versus the population, you would have to make the assumption that only 8,079 people are in the town at any given time.”

From late August, adding the population of Longwood adds to the numbers (a community of about 13,000) to May. Add in the numbers from Hampden-Sydney College plus factoring that Farmville is also a magnet for communities surrounding the town and Davis figures the true population from August to May is “much higher than 8,079.”

Davis looks at crime number comparisons — the number of larcenies that occurred last June, for example, versus the number in the same month this year. If they realize more this year than last, they look at causation, what are the factors that contributed to the increase. “And so that’s how we do analytical policing,” he said.

He looks at assaults (even including minor assaults as part of the violent crimes stats because they can lead to greater violence), looks at a 4-5 year average, how they trend, what happened in the particular years (such as economic factors or a specific incident) and tries to identify anomalies within the data and explain them for a clearer picture.

He also looks for trends, increases that can be signs of an unknown problem. The analysis can also identify the types of problems in a particular area.

Just looking at the figures, however, can give a skewed view. If shoplifting becomes a police department’s focus, the arrests and reports could be higher, and that could, in turn, indicate an increase in an individual’s chance in becoming a victim.

It wouldn’t mean more things were stolen, but effectively, the more aggressive the police department, the worse the crime statistic would look.

“If you really wanted … numbers to drop, you quit enforcing shoplifting, for example,” Davis said, citing the ridiculousness of such figures. “You just let the insurance companies take care of it, and then everybody feels safer because our numbers are lower.”

Davis suggests talking to neighbors in the community where one is planning to move.

“What I think would be [one] the most important aspects of considering the safety level … is to speak with residents, business owners within a community, contact the local police department and inquire about concerns and issues that you have related to crime in the area and then evaluate that for yourself, not allowing the data to simply speak for itself,” Davis said.

(The town of Farmville has statistics dating to 2009 at farmvilleva.com/crime-statistics-. Incident statistics could include cases that that could not be proved nor clearly disproved.)