undefined | Better HN

0 pointsXylakant12y ago0 comments

You're glossing over my main point: Picking a single number without context or interpretation from a statistic is devoid of any meaning.

The statistic measures "reported crimes" as a total value, not even per capita. That's a very long shot from "committed crimes". It's really just only what people report, if it's a crime or not. The number of reported crimes is most likely below the number of actual crimes for most crimes. It's a statistic which must be regarded with big caution. It's an indicator and as such has a value, but without context and analysis the number itself is basically meaningless. Take for example "fraud with help of internet" - that's a crime that basically didn't exist in 2003. It probably wasn't codified in law until much later and it's one of the fastest growing categories. Such changes skew the statistics, context is important. Even the reverse in trend may be due to external factors such as less people reporting drug abuse as a crime.

0 comments

draugadrotten12y ago

I agree that crime statistics are highly context sensitive and must be handled with caution. Sweden's rape statistics are a good example. Sweden has more rapes per capita than most countries, it's even the second highest in the world. However, that's because Sweden's definition of rape is including so much more than any other country. Ask Julian Assange about it.

http://en.wikipedia.org/wiki/Rape_statistics#Sweden

That new crimes are codified in law, such as "internet fraud", and is one of the fastest growing seems to causing more crimes, not fewer. It doesn't invalidate the comparison between years for all crimes, and unless you want to argue that newly codified crimes are less serious than old ones, I can't really see a problem with new laws being included in statistics, too.

I'd like to conclude saying that I appreciate your posts on this subject and have read them with interest.

idupree12y ago

Wikipedia's citation in that section[1] says that Swedes are more likely to report crimes, but have a similar actual amount of crimes.

"Sweden stands out within the entire area of crimes against the person in particular, because the registration of crime is more extensive than in the majority of other countries in Europe. This forms the background to, for example, the fact that ten times as many cases of assault are registered in Sweden as in Greece."

"But in the victim survey that exists that permits basic comparison between levels of exposure to sex crimes in ten European countries, Sweden does not top the list, as in the reporting statistics. Instead, Sweden is around the average mark, which is also the case for assaults and threats, despite the fact that compared to other countries, we have many such crimes reported. However, the fact that we are not any worse than many other countries does not mean that the situation is good. If any rape is committed, the level is too high."

[1] http://www.bra.se/bra/bra-in-english/home/news-from-bra/arch...

XylakantOP12y ago

That new crimes are codified in law, such as "internet fraud", and is one of the fastest growing seems to causing more crimes, not fewer. It doesn't invalidate the comparison between years for all crimes, and unless you want to argue that newly codified crimes are less serious than old ones,

There is no problem per se in including newly codified crimes in statistics. It's also perfectly fine that that this causes a rise in those statistics. It still compels caution when reading those statistics:

Let's best take a totally contrived example: Beginning with the first of December, eating Surströmming is forbidden due to new environmental protection laws. This will cause a sudden spike in the crime statistics because once acceptable behavior is now outlawed. It also causes the police to suddenly solve a higher percentage of crimes since the perpetrators are easily identifiable by their smell. Did the world get better? Worse? The statistic just tells us a naked number.

Something similar happens with fraud with help of the internet and probably more with illegal access to computer systems: People were doing it before, but it never showed up in the statistics since even though it was still as bad of a behavior as it is now, it was not illegal since there was no appropriate law.

Figures for rape and sexual molestation are rising? What does that mean? More people getting raped? Or does it mean that even though less people are raped, a higher percentage are in the position to report it? The definition of rape has - rightfully - been broadened in the last decade. In germany, rape in a marriage was only codified as "rape" in 1997 (!) (in words: nineteen nintyseven). That certainly causes a rise in the relevant statistics, but it doesn't mean that more women are raped - it just means that visibility has increased. (which, by the way, is a good thing in itself)

That's what I mean by "context matters". You can't derive any meaning from the naked numbers. In the extreme case, a rise in crime stats may "just" mean that the society decided to outlaw some specific behavior. In reality it's probably a mixed bag of a lot of reasons.

EDIT: Formatting for the quote.

berntb12y ago

>>However, [the statistics are the 2nd highest on the planet] because Sweden's definition of rape is including so much more than any other country.

I've seen that claim often, but never some real facts which compares data. You don't have any, either?

j / k navigate · click thread line to collapse