Stone Soup

It has been a while since the issue of place rating has come to the fore, at least in any innovative way which required comment but, obviously, that has changed. A news report, wait, a series of news reports derived from one particular not totally disinterested source has been making the rounds throughout the local media and is being used to hype the “The City Has Gone Crazy” narrative.

It all starts with a website which is clearly trying to enter the home security business as an agent and a vetting service for alarm systems. Their approach is clever. Rather than rating and comparing cities nationwide, where at most say…10 cities can be on the 10 most dangerous list, they go state by state making “Top 10”, “Top 5”, or even “Top 20” lists for each state thus putting around 300 municipalities on alert for being unusually dangerous.

Of course, the source of this information is an FBI data set which has attached to it these words from a page labeled “Caution Against Ranking”:

The UCR Program provides a nationwide view of crime based on statistics contributed by local, county, state, tribal, and federal law enforcement agencies. Population size is the only correlate of crime presented in this publication. Although many of the listed factors equally affect the crime of a particular area, the UCR Program makes no attempt to relate them to the data presented. The data user is, therefore, cautioned against comparing statistical data of individual reporting units from cities, counties, metropolitan areas, states, or colleges or universities solely on the basis on their population coverage or student enrollment. Until data users examine all the variables that affect crime in a town, city, county, state, region, or other jurisdiction, they can make no meaningful comparisons. (emphasis mine)

I have written page after page of explanation about rankings systems which use this data to make meaningless comparisons. Rather than reiterate it here, I will provide this link.

What is unfortunate is that one local news outlet in particular has decided to run multiple stories based on this faulty data AND has managed to add to the misinformation both by misattributing the rankings directly to the FBI, and by eliciting comments from uninformed citizens and reporting their opinions as fact. WWLP first runs reports based on this faulty data on Tuesday, October 7th and then follows up with more reports on multiple newscasts for three consecutive days. I make that somewhere between 5 and 9 broadcasts of a similar package or story. Then on Sunday the 12th of October they run this story which not only makes the attribution mistake but which then quotes a citizen as saying that:

“It doesn’t surprise me. Springfield unfortunately has gone down hill in the past 20 years or so. I’ve seen a lot more violence…”

Only the data shows a 65% percent drop in violent crime from 20 years ago in the city of Springfield which, by the way, is an appropriate use of the data as it is prepared specifically for cities to compare their own progress in fighting crime. The reporter gets away with allowing this assertion, of course, because this particular Springfield resident did say what he said, as inaccurate as his perception may be. I would go one step further and say that when the man says

“I’ve seen a lot more violence, a lot more gang activity”

he has “seen it” on the news. I am sure I have seen gang activity. I am even more sure that I did not notice it as such. What does “gang activity” look like? Is it this:

or this?

or this?

Is it anytime more than one black, Hispanic, or mysteriously hooded young man walks down the street?

Compare this obviously exploitative treatment of “meaningless” data to the coverage of real data from an objective source: a system which locates and tabulates the number of gun shots in certain parts of the city. When the Shotspotter system employed by the city detected 60% fewer gunshots during a sixth month period in a year over year comparison this same news outlet ran one news package which aired twice and contained no “man on the street” interviews and was never followed up. It went unmentioned that the 60% drop in gunshots corresponded to a 40% drop in murders over the same six month period, that the city looks to be in line for a 20% drop in murders for the year, or that overall crime is down by 2/3 over the last 20 years.

(Columns are: Year/Months Reporting/Population/Total Violent Crime)

It also went unmentioned that, if one doesn’t mind going “meta” in the area of bullshit, that the ranking of Springfield as only the 5th most dangerous city in the Baystate would constitute a drop of up to four spots because, as recently as 2012, Springfield was listed as the 12th most dangerous city nation wide and the most dangerous in Massachusetts…that from a DIFFERENT company which also uses the data you’re not supposed to use to compare to compare oranges to apple orchards.