Suspicious Crime Statistics

[Amy]

Hi,

I'm not sure if this is the right forum for this, but its the only statistics related forum that I could find.

I am in the process of putting together a report on crime and sustainability, and one of the indicators used is the murder rate. However, I am quite suspicious of the official statistics put out by the police services.

Please see the attached Excel File for the graph showing the trend from 1994 to 2006. You can see that there is a very sudden and sustained drop in the murder rate in 03/04. Superficially that is great news, but I can't help wondering whether or not the data has been falsified.

These numbers are for a large city. There has been no obvious external factor, such as redrawing of municipal boundaries, that could have caused this.

So, I just wanted to ask people's opinion on the matter: Do you think that this could be a real decrease? What kind of mechanism could be responsible for such a sudden decrease in a particular crime statistic? There have not been similar sudden decreases or increases in other violent crimes. All the other data (for other crime categories) that I have shows a gradual increase, gradual decrease or no change, as I would have expected.

Thanks for any advice.

 

[CMSed]

Amy,

Falsified may be a bit strong. When I see a change this big with no obvious explanation, the first question that I'd ask is if there was a change in the way that this department collected its data from 02/03 to 03/04. Many of the problems that I've dealt with in police call-for-service or incident data have some instrumentation issue at their root.

Chris

 

[Amy]

Well, I wouldn't put it past the government to blatantly falsify these stats, but I will give them the benefit of the doubt.

I thought about a change in methods of data collection, but how exactly would that change for murder? Its not something that can be redefined or split into different categories. There also haven't been any similar changes in any of the other crime stats, including armed robbery, rape, theft etc. I haven't checked the stats for attempted murder - I will do that when I get back to work tomorrow.

Thanks for the reply though.

Any other ideas, anyone?

 

[CMSed]

Not necessarily definitional changes - but changes in the recordkeeping protocols themselves can cause this as well. While I was in grad school, I worked for a large city PD. During that time, the computer-assisted dispatch database software was replaced and the departmental ten-codes were revised as well. The combination of the two caused the kind of phenomenon that you're seeing in this data. Until we reconciled those changes with our databases, the data was a total mess.

 

[Amy]

Thanks for your reply.

I have a feeling I'm actually going to have to approach the Police Services and try to get to the bottom of this - if it was some kind of procedural change that would be the only way to find out.

*sigh* They are notoriously unco-operative, so I'm not confident that I will ever find out what happened.

Thanks for your help though