International news media seems to think that the Swedish city of Malmö is being overrun by gun violence. In reality, the definition of a shooting differs between Malmö and the other two main cities. In Malmö, it does not have to be an actual gun that was fired and there is no requirement for forensic or eyewitness testimony. The geographical areas also differ, with Malmö only covering the main city and the urban area Arlöv. Once researchers looked through the data and counted the number of actual shootings, the figures dropped by 60-75%. The Malmö police does offer some justifications of their classification scheme, but in the end, organizations that gather statistics have an intellectual responsibility to ensure that their data are not easily abuse by being clear with definitions and what can and cannot be inferred from their data.
A recent analysis by Swedish Council for Crime Prevention was covered in the newspaper Sydsvenskan. The sad reality is that fear propaganda gets front page news, whereas a careful statistical analysis gets only a small notice in most papers.
What is a shooting event?
Perhaps surprisingly, the definition of a shooting differs drastically between the three largest cities in Sweden (Stockholm, Göteborg and Malmö). In Stockholm and Göteborg, a shooting requires two criteria: (1) the discharge of a gunpowder-loaded projectile and (2) corroborating forensic evidence or eyewitness (or earwitness) testimony. In Malmö, a lot of other things are also classified as a shooting, such as firing of airsoft guns, slingshots, damages to windows that look like a bullet hole and even damages done by stone chips. Thus, the figures cannot be naively compared because they do not measure the same things.
What geographical area is considered in the different cities?
Another major difference is that the geographical area that is considered differs substantially. For Stockholm, it is the entire county. For Göteborg, it is Storgöteborg, which is the city including all surrounding municipalities. In contrast, for Malmö, it is just Malmö city and the urban area Arlöv. This means that the denominator is inflated for Stockholm and Göteborg and so the figures for Malmö for shooting per area will be artificially inflated.
How many reported shootings in each city?
So let us look at some data. How many reported shootings were there in each area? This data comes from the BRÅ rapport Shootings 2006 and 2014: Extent, Spread and Damages.
|Malmö city + Arlöv||140||97||114||122|
But far from all of those shootings in Malmö were actual shootings. In fact, a majority were not.
How many of the 114 reported shootings in Malmö were real shootings?
It is of course always difficult to tell if something is a real shooting or not. To get around this, researchers at the Swedish Council for Crime Prevention (BRÅ) decided to find out how many of the reported cases of shootings in Malmö would have been classified as a shooting had it occurred in the other two areas.
The answer? It dropped form 114 to 29, a loss of almost 75%. Could this be just a fluke? The researchers also looked at data from 2014. There the official figure was 122 shootings, but only 47 of them would have been considered shootings had they occurred in Stockholm or Göteborg. This means that over 60% of reported shootings in Malmö during 2014 were not actual shootings. The researchers at BRÅ still thinks that there are more shootings in Malmö than the other two cities, but nowhere near as much as the official statistics claim.
Why is there a discrepancy of shooting definitions?
The definition of a shooting in Malmö was changed in the early 2010s due to a serial shooter called Peter Mangs. Known as the New Laser Man, he stalked and shot several immigrants in Malmö in the early 2010s using a laser sight that he targeted his victims with prior to shooting them. The Laser Man name comes from John Ausonius, a previous serial shooter in Stockholm from the early 90s. In Malmö, Mangs often did not leave any forensic evidence after him and there were several cases where others had used airsoft weapons against public transports, so they scrapped most of their requirements for something to be reported as a shooting event.
How does the Malmö police justify their classification scheme?
They offer up a number of explanations: they wanted to include all kinds of shooting weapons, giving the police an overview of the problem, and because the average person cannot tell the difference between real gun and an airsoft gun. They do, however, acknowledge that there have been a lot of mischaracterizations when their figures have gained mass media attention.
Intellectual responsibility of statistics-gathering organizations
It is important for organizations that gather statistics, whether government or independent, to understand that their data can, and perhaps will, be abused for nefarious goals. It is no longer enough to just state the data. The data has to be clearly defined, explained and put into context. The definitions of what is being measured has to appear prominently in any publication and not just a sentence or a footnote here and there. It is important to state clearly what and how conclusions can be reached by the data, but even more important is to explain what conclusions cannot be drawn from the data and why.
Categories: Debunking Misuse of Statistics