Crime statistics true or not

[titus75]

I read all the time about the crime rate dropping, murders, robberies, serious felonies all supposedly down. My perception, and it seems to be quite a few other peoples perception also, is that these stats are either simply wrong, being misinterpreted or "juked" so to speak.

http://oag.ca.gov/crime

http://www.fbi.gov/about-us/cjis/ucr/ucr

http://bjs.ojp.usdoj.gov/

[Dreaded Claymore]

Try again?

I'm not sure what your post is asking.

If you think you're being falsely informed, then you'd want to find out the real answer for yourself. The thing is, you'd do that by collecting data and performing a statistical analysis. That'd be really hard. You could look at statistical data that someone else has compiled, except that the problem in the first place is that you don't trust it.

Looks like you're stuck. Too bad.

[Librarian]

Why do you distrust the Uniform Crime Reports?

There are places whose reporting is missing, and some that classify crimes differently than directed, but UCR is the best crime reporting system we have.

I'm personally rather skeptical of the British data...

[titus75]

Exactly, I'd like some opinions, perhaps some experienced ones, as to how these statistics are compiled and sorted, maybe some LEO opinion.

[titus75]

Why do you distrust the Uniform Crime Reports?

There are places whose reporting is missing, and some that classify crimes differently than directed, but UCR is the best crime reporting system we have.

I'm personally rather skeptical of the British data...

So you take the UCR for the truth? You believe crime is down?, I'm just asking your opinion. I guess I mistrust it because it's compiled from local police departments and sheriffs offices who are under pressure to show the taxpayers a supposed drop in crime. Also the info in the UCR is voluntarily submitted, I think that makes a big difference.

[jwkincal]

The FBI collects information that is reported by local agencies. I expect that the compilation is statistically sound, but the reporting will tend to be hit-and miss as we are talking about thousands of jurisdictions, some huge and some very small, some well-staffed and some not so much, some diligent and some slovenly...

Bottom line, the data probably indicates accurate trending, but the absolute values in terms of the number and types of crime is subject to common statistical and reporting errors.

I'm not in LE but I do know how this stuff works from a market research perspective; this data is essentially demographic crime analysis... it passes through the hands of many fallible humans before it is assigned the sacred value of "statistical data." Use at your own risk.

[G60]

Don't you think local law enforcement would rather report an increase in crime to justify their budgets in a time when they're under intense pressure to cut them?

When crime increases, the general public doesn't ask "what is LE doing with my tax dollars?", they say "we need more officers on our streets!"

[titus75]

Yes and no, a local sheriff or police chief who has a high crime rate will be out of a job pretty quickly, it's in most high ranking law enforcement officials, not to mention politicians, best interests to show a drop in crime, especially violent crime. Funding is based largely on population i.e. tax base, with crime being only one of the factors.

[Librarian]

So you take the UCR for the truth? You believe crime is down?, I'm just asking your opinion. I guess I mistrust it because it's compiled from local police departments and sheriffs offices who are under pressure to show the taxpayers a supposed drop in crime. Also the info in the UCR is voluntarily submitted, I think that makes a big difference.

Truth?

Well, no; statistics are not 'truth'.

I think they (UCR) devote sufficient resources to checking the data - see the doc UCR Methods of Quality Control - that this is a pretty good instrument for what it is.

UCR doesn't seem to map very closely to the NCVS (National Crime Victimization Survey); see this doc: http://bjs.ojp.usdoj.gov/index.cfm?ty=pbdetail&iid=802

But as a systemic problem, I don't perceive either survey as 'cooked'.

I do think that problem is institutional with the British data from about 1990 forward. They've redefined crimes, redefined their reporting year, and (IMO and in the opinion of Joyce Lee Malcom at George Mason U) applied their laws more harshly against what seem to be victims than what seem to be the criminals.

[tyrist]

The statistics are about as accurate as a human being can manage. If you wanted to see manipulation it would be best to look into how the various departments list crime (eg. Petty theft vs burglary).

[titus75]

Truth?

Well, no; statistics are not 'truth'.

I think they (UCR) devote sufficient resources to checking the data - see the doc UCR Methods of Quality Control - that this is a pretty good instrument for what it is.

UCR doesn't seem to map very closely to the NCVS (National Crime Victimization Survey); see this doc: http://bjs.ojp.usdoj.gov/index.cfm?ty=pbdetail&iid=802

But as a systemic problem, I don't perceive either survey as 'cooked'.

I do think that problem is institutional with the British data from about 1990 forward. They've redefined crimes, redefined their reporting year, and (IMO and in the opinion of Joyce Lee Malcom at George Mason U) applied their laws more harshly against what seem to be victims than what seem to be the criminals.

You have me there, that was a poor choice of a word. I suppose I'm thinking lies, damn lies and, of course, statistics. I wasn't aware of the British manipulation of crime stats.

[zhyla]

My perception

Which do you think is more likely, that your perception is off by +/- 50%, or that the crime stats are inaccurate by +/- 1%? People are horrible judges of large numbers unless we actually write everything down and add it all up.

And don't forget changes in what gets reported. On a slow day all you read about is stabbings. But yesterday that got pre-empted by the cops shooting some nutjob around here. Numbers don't lie.

[titus75]

Which do you think is more likely, that your perception is off by +/- 50%, or that the crime stats are inaccurate by +/- 1%? People are horrible judges of large numbers unless we actually write everything down and add it all up.

And don't forget changes in what gets reported. On a slow day all you read about is stabbings. But yesterday that got pre-empted by the cops shooting some nutjob around here. Numbers don't lie.

Again, a poor choice of a word, call it my belief that the statistics are not accurate in that the overall crime rate, in particiular serious felonies, are underreported or misreported as lesser offences, numbers don't lie but people do.

[spalterego]

The problem is the difference between perception and reality. Your perception of the crime rate is largely driven by news. The news media have become more and more fixated on crime stories. When I was growing up, the national news hardly ever reported on crime and my local news only reported on truly significant crime. I have changed locations from where I grew up but my perception is that all news channels have abandoned hard news stories (regardless of whether it is a "right" or "left" issue) and a large portion of the news follows the mantra, "if it bleeds it leads".

While the UCR are certainly fallible and individual agencies/officers may try to misrepresent crime the really "hard" statistics such as homicide show a downward trend that is in line with the general overall downward trend indicated by the "softer" categories.

[titus75]

Good and interesting opinions. Maybe I'm a just a misanthrope or you're all a bunch of cockeyed optimists.