There is no single statistical source for making a valid direct comparison between cities, and it is indeed possible that Houston has a crime rate similar to NYC, although there would be no comparison using total number of crimes - NYC would be far and away much higher. And yes, the tendency to underreport crimes in high crime areas is true virtually everywhere.G26ster wrote:So what is the basis (reliable statistics) for deciding/saying that NYC has a higher crime rate than Houston, or any other city?Excaliber wrote:Total crimes and crime rate are different concepts.G26ster wrote:Based on the thread title, I'm a bit confused. Every crime rate calculator I look at shows Houston, TX has nearly twice the violent crime rate per 100k as NYC, and NYC has 4 times the population of Houston. What am I missing?
Total crimes is the total number of crimes reported. Crime rate is the number of crimes per 100,000 persons in a given population.
Both numbers are influenced by the rate at which crimes are reported (in many rough neighborhoods where people have little confidence many crimes will be investigated or solved, they report in low numbers) and the consistency with which they are categorized by the local law enforcement agency.
These and other factors are why the FBI cautions not to use the UCR statistics as direct comparisons between cities.
Wouldn't this be true for both NYC and Houston?Excaliber wrote:(in many rough neighborhoods where people have little confidence many crimes will be investigated or solved, they report in low numbers)
The rates for the 7 crimes tracked by the FBI in their UCR reports are not even across the board. For example, two cities could have an identical rate of total crimes per 100K population, but one of those might have an extremely high larceny rate but a very low homicide rate, and the other city the exact opposite. The nature of the crimes above the national average rates is important in assessing relative safety.
There are other factors in play as well, and risk managers often use more refined tools than the UCR to look at the risk of background crime in a given area. One of the most prominent of these is the CrimeCast report prepared for specific locations. In addition to reported crimes, it takes into account demographics, income levels, insurance company loss statistics, and victimization surveys of unreported crimes to deliver an assessment of crime risk (likelihood and impact) in a given area broken down by census tract. It is much more useful for understanding relative risk than the raw UCR statistics.