A fairly simple question to follow on in the vein of media criticism in which I have placed this blog in recent months: if the crime rate in my city had gone up overall by 14% including a 40% increase in murders, would the article elaborating on the release of FBI data revealing that information have been given a headline which included the idea that “crime” was “up” in “Springfield”? There are two possible answers; “yes”, and “you don’t understand how the media works”.
When the most recent release of FBI data showed that overall crime was down 14% in Springfield, including a 40% drop in murders we were given this headline:
No mention of Springfield whatsoever in the headline.
And this nuanced take within the article on the extreme drop in the number of murders in Springfield:
I find it comical that the admonition to take care with respect to the apparently enormous drop in the number of murders was in part expressed by stating that the remarkably low number was so low that it shouldn’t be taken as any sort of positive indicator, but rather as a statistical outlier of sorts:
Imagine the reverse: “Well, yes, the number of murders was spectacularly high last year but that it was so much higher than the extremely low number of the previous year is more indicative of a statistical anomaly than of any problem with violence.”
That article would never see the light of day.
For anyone new to the topic let me address the basics when it comes to FBI crime statistics. First of all, despite an honest attempt at calibration by the FBI, different jurisdictions have both different criteria for crimes and see greatly differing levels of reportage. The bottom line here is that, in many cases, places which experience more crime and/or have police forces viewed as either incompetent or hostile to its citizenry will benefit statistically from the fact that crimes are less likely to be reported.
Another enormous variant is geographical. What is considered a city, for example, varies greatly from region to region. Comparing Jacksonville, Florida with its population of 800,000 people living within a boundary of over 900 square miles, to a city like Springfield where a 900 square mile territory would include a population of nearly the same size, but which instead occupies only 30 square miles and has 155,000 of population creates a circumstance where the crime endemic to poorer inner city neighborhoods is amplified in the smaller geographical community and watered down in the more sprawling metropolis.
For these reasons the FBI report itself warns against using the data to compare community to community but rather encourages each community to use the data to analyze its own progress relative to the reduction of crime.
To move beyond that, if one does wish to use crime data to compare community to community then the best data is the number of murders per capita in each community. This is because both the definition of murder and its reportage are very consistent nation wide. However, as the Masslive article indicates, the nature of murder, having a relatively small number of occurrences when compared to the overall number of violent crimes, is such that it should be used in multiple sets of years, perhaps 4 or 5 at a minimum, so that the statistical anomalies of small numbers can best be overcome.
If there is one thing that you can count on with the local media it is that their ability to express nuance will magically appear when there is good news about Springfield and crime.
Pingback: Rational Urbanism | This is also what democracy looks like