Crime Rate Statistics
September 29, 2011 3:26 PM Subscribe
Please clear up the use of statistics by two people trying to compare crime rates in one huge city (NYC) to Charleston, SC. a much smaller city.
On a travel board I frequent, two people got into an argument about the crime rates in their respective cities. The New York poster put up a bunch of charts showing that the crime rate per hundred thousand people was lower in New York City than in Charleston, SC. The SC responder took umbrage and attacked the NY guy's methods. He (the SC poster) offered the following garbled explanation of why the two cities couldn't be compared in this way, i.e. by using number of crimes per one hundred thousand people:
Ok, since I need to explain statistics again, those numbers are based on acts/100K people, therefore with NYC have 8.4 million vs 384 thousand, 10 violent acts a year here will give you a violent crime rate of 3/100k where as the same 10 acts will give you a crime rate of .11/100k so another words in order to have the same crime rate in NYC that Charleston has with 10 acts, NYC needs to have 840....
So tell me where is there more crime?
Can anyone figure out the argument the SC guy is making? If so, is he correct in implying that calculating crimes per one hundred thousand people isn't an accurate way to compare two such different cities?
On a travel board I frequent, two people got into an argument about the crime rates in their respective cities. The New York poster put up a bunch of charts showing that the crime rate per hundred thousand people was lower in New York City than in Charleston, SC. The SC responder took umbrage and attacked the NY guy's methods. He (the SC poster) offered the following garbled explanation of why the two cities couldn't be compared in this way, i.e. by using number of crimes per one hundred thousand people:
Ok, since I need to explain statistics again, those numbers are based on acts/100K people, therefore with NYC have 8.4 million vs 384 thousand, 10 violent acts a year here will give you a violent crime rate of 3/100k where as the same 10 acts will give you a crime rate of .11/100k so another words in order to have the same crime rate in NYC that Charleston has with 10 acts, NYC needs to have 840....
So tell me where is there more crime?
Can anyone figure out the argument the SC guy is making? If so, is he correct in implying that calculating crimes per one hundred thousand people isn't an accurate way to compare two such different cities?
Also the condescending tone of that comment was directed at him, not you!
posted by p3on at 3:31 PM on September 29, 2011
posted by p3on at 3:31 PM on September 29, 2011
Bull.
acts/100k is a perfectly valid metric, because what you're really getting at is the probability of becoming a victim. If there are 100 robberies in a city of 8 million, there's a 1 in 80,000 chance you'll be the victim. If there are 100 robberies in a city of 400,000, there's a 1 in 4,000 chance you'll be the victim.
posted by Oktober at 3:32 PM on September 29, 2011
acts/100k is a perfectly valid metric, because what you're really getting at is the probability of becoming a victim. If there are 100 robberies in a city of 8 million, there's a 1 in 80,000 chance you'll be the victim. If there are 100 robberies in a city of 400,000, there's a 1 in 4,000 chance you'll be the victim.
posted by Oktober at 3:32 PM on September 29, 2011
I'm confused as well. A rate is a rate - last I checked, 100,000 was the same number in NYC as it was in Charleston. There are fewer units of 100,000 in Charleston, sure, but crimes per 100,000 people is a reasonable way to evaluate the safety of a city.
posted by pdb at 3:33 PM on September 29, 2011
posted by pdb at 3:33 PM on September 29, 2011
I had this argument with some friends of mine. I said I felt safer in a larger nearby city than in our hometown, and cited the relevant crime rates. I was considered an idiot by that group because "there are fewer murders, even if the rate is higher." Hmm.
posted by sonic meat machine at 3:36 PM on September 29, 2011 [1 favorite]
posted by sonic meat machine at 3:36 PM on September 29, 2011 [1 favorite]
World Famous, that would completely depend on neighborhood. I'm sure there are some blocks in New York that are way more dangerous than Charleston and equally sure there are many that are thousands of times safer.
posted by Maias at 3:51 PM on September 29, 2011
posted by Maias at 3:51 PM on September 29, 2011
Can anyone figure out the argument the SC guy is making?
The argument he is making is that there is "more crime" in New York City than Charleston on an absolute basis
If so, is he correct in implying that calculating crimes per one hundred thousand people isn't an accurate way to compare two such different cities?
Per capita is a great way of comparing crime rates between areas of different populations. Murders per geographic area would be another way, but that would be highly skewed by population density. Ultimately though what's happening is that some people refuse to believe how safe New York City really is, and that they are much more likely to be killed by their local friends and neighbors.
posted by 2bucksplus at 3:52 PM on September 29, 2011 [2 favorites]
The argument he is making is that there is "more crime" in New York City than Charleston on an absolute basis
If so, is he correct in implying that calculating crimes per one hundred thousand people isn't an accurate way to compare two such different cities?
Per capita is a great way of comparing crime rates between areas of different populations. Murders per geographic area would be another way, but that would be highly skewed by population density. Ultimately though what's happening is that some people refuse to believe how safe New York City really is, and that they are much more likely to be killed by their local friends and neighbors.
posted by 2bucksplus at 3:52 PM on September 29, 2011 [2 favorites]
so another words in order to have the same crime rate in NYC that Charleston has with 10 acts, NYC needs to have 840
Also, um, actually, the ratio is 21.9, so ten acts, 219 (rounding).
Sometimes, there's no arguing with crazy. As a stats guy, I could make no sense of his math or logic.
posted by General Malaise at 3:58 PM on September 29, 2011
Also, um, actually, the ratio is 21.9, so ten acts, 219 (rounding).
Sometimes, there's no arguing with crazy. As a stats guy, I could make no sense of his math or logic.
posted by General Malaise at 3:58 PM on September 29, 2011
"So tell me where is there more crime?"
But he (Charleston guy) is right, of course. There is "more" crime in NYC than in Charleston. Just as there is more in South Carolina than in Charleston.
I think what you say to the guy is, "If I walk passed a thousand New Yorkers, and you walk passed a thousand Charlestoners, who will have walked passed the most people who have committed violent crimes? Even if it takes you twelve times as long to find those thousand, you're going to have seen the most." (actually by quite a bit)
But it want work, regional jingoism can't be overcome, and we have lots of people around who believe that repetition creates truth.
posted by Webnym at 4:11 PM on September 29, 2011
But he (Charleston guy) is right, of course. There is "more" crime in NYC than in Charleston. Just as there is more in South Carolina than in Charleston.
I think what you say to the guy is, "If I walk passed a thousand New Yorkers, and you walk passed a thousand Charlestoners, who will have walked passed the most people who have committed violent crimes? Even if it takes you twelve times as long to find those thousand, you're going to have seen the most." (actually by quite a bit)
But it want work, regional jingoism can't be overcome, and we have lots of people around who believe that repetition creates truth.
posted by Webnym at 4:11 PM on September 29, 2011
Like Webnym says, I think regional jingoism underlies this. But he's trying to use statistics when it is clear he doesn't understand grammar-school math (and is being smug about it). It's as if he's trying to normalize by the numerator instead of the denominator.
Now, if you want to reject the idea of crime rate as a valid measurement, as World Famous is doing, that's one thing. I'll disagree, but it's a separate issue. The SC person in the original question is not rejecting crime rates. He's completely misunderstanding what they mean. Perhaps intentionally.
posted by adamrice at 4:17 PM on September 29, 2011 [1 favorite]
Now, if you want to reject the idea of crime rate as a valid measurement, as World Famous is doing, that's one thing. I'll disagree, but it's a separate issue. The SC person in the original question is not rejecting crime rates. He's completely misunderstanding what they mean. Perhaps intentionally.
posted by adamrice at 4:17 PM on September 29, 2011 [1 favorite]
Presumably there's more crime in NYC, but the average person is less likely to be affected by it, hence the ratio of number of crimes per capita to get a crime rate.
I think his argument is trying to be the first part of that sentence, but so what? It has nothing to do with which place is safer by the numbers. His language and logic is so garbled that he just sounds like a doofus.
Number of crimes does not equal "crime rate". Rate implies ratio when you're talking about statistics.
posted by oneirodynia at 4:38 PM on September 29, 2011
I think his argument is trying to be the first part of that sentence, but so what? It has nothing to do with which place is safer by the numbers. His language and logic is so garbled that he just sounds like a doofus.
Number of crimes does not equal "crime rate". Rate implies ratio when you're talking about statistics.
posted by oneirodynia at 4:38 PM on September 29, 2011
The SC is right - according to his own metric (So tell me where is there more crime?). NYC does indeed have more crime - as an absolute figure - than SC.
He is confusing the meaning of a crime rate though - though SC has less crime (according to those numbers), each SC resident has a higher probability of being a victim of said crime.
posted by twirlypen at 4:39 PM on September 29, 2011
He is confusing the meaning of a crime rate though - though SC has less crime (according to those numbers), each SC resident has a higher probability of being a victim of said crime.
posted by twirlypen at 4:39 PM on September 29, 2011
Okay, I've been going over and over this for a few minutes now, and I think I've come up with the only plausible explanation of his "logic."
Math aside (his math is terrible), what it seems to be is: Look at the crime rates: 10 murders in SC makes the rate .00002604. 10 murders in NYC and it's .00000119. NYC has such a lower crime rate then! But we all know there is more crime in New York! We know it, dangit! So, crime per capita is useless!
It reminds me of some who like to play that the "probability" of life coming about through evolution is so low that it's impossible, therefore creation! It's (really bad) circular logic.
posted by General Malaise at 4:49 PM on September 29, 2011
Math aside (his math is terrible), what it seems to be is: Look at the crime rates: 10 murders in SC makes the rate .00002604. 10 murders in NYC and it's .00000119. NYC has such a lower crime rate then! But we all know there is more crime in New York! We know it, dangit! So, crime per capita is useless!
It reminds me of some who like to play that the "probability" of life coming about through evolution is so low that it's impossible, therefore creation! It's (really bad) circular logic.
posted by General Malaise at 4:49 PM on September 29, 2011
is he correct in implying that calculating crimes per one hundred thousand people isn't an accurate way to compare two such different cities?
Crimes per capita is the only way to compare crime rates. There is no other accepted way. If you look up Crime rates of cities or regional crime rates, they are listed by "cases per 100,000 people".
posted by deanc at 5:22 PM on September 29, 2011
Crimes per capita is the only way to compare crime rates. There is no other accepted way. If you look up Crime rates of cities or regional crime rates, they are listed by "cases per 100,000 people".
posted by deanc at 5:22 PM on September 29, 2011
Comparing crime rates properly is extremely tricky. First, the absolute number of murders in Charleston is tiny, so you really need to do a multi-year average to avoid any annual fluctuations mucking things up. Second, you really want to look at an entire metro area, not just arbitrary city limits -- in Charleston less than 1/5th of the metro population lives within the city limits, and in NYC it's less than half.
(Looking at city limits usually gives you nonsense numbers because a lot of people are working, shopping, and being victims of crimes while in the city, without actually living there. They increase the numerator of the crime rate formula without increasing the denominator.)
The poster definitely didn't bring either of these points though.
posted by miyabo at 6:37 PM on September 29, 2011
(Looking at city limits usually gives you nonsense numbers because a lot of people are working, shopping, and being victims of crimes while in the city, without actually living there. They increase the numerator of the crime rate formula without increasing the denominator.)
The poster definitely didn't bring either of these points though.
posted by miyabo at 6:37 PM on September 29, 2011
His number for the population of Charleston is completely off, its population is only around 108,000.
posted by mareli at 7:08 PM on September 29, 2011
posted by mareli at 7:08 PM on September 29, 2011
Can anyone figure out the argument the SC guy is making?
No insult intended to TWF, but I think you're being too generous here.
He is saying "I live in South Carolina and it is AWESOME and GOOD and not HORRIBLE AWFUL like New York. New York sucks because it has more total violent crimes than Charleston does."
"If so, is he correct in implying that calculating crimes per one hundred thousand people isn't an accurate way to compare two such different cities?"
No, it's about the only reasonable or sane way to do so. He's just wrong, and his proposal is goofy beyond measure.
He's just het-up about a perceived insult to his city or state, and being a bit of a ninny trying to defend it.
posted by ROU_Xenophobe at 8:18 PM on September 29, 2011
No insult intended to TWF, but I think you're being too generous here.
He is saying "I live in South Carolina and it is AWESOME and GOOD and not HORRIBLE AWFUL like New York. New York sucks because it has more total violent crimes than Charleston does."
"If so, is he correct in implying that calculating crimes per one hundred thousand people isn't an accurate way to compare two such different cities?"
No, it's about the only reasonable or sane way to do so. He's just wrong, and his proposal is goofy beyond measure.
He's just het-up about a perceived insult to his city or state, and being a bit of a ninny trying to defend it.
posted by ROU_Xenophobe at 8:18 PM on September 29, 2011
One way to think about these things is to see if they make sense in extreme cases.
Say Charleston had 384,000 murders, and New York had 385,000. NYC stills have more (absolute) crime. Does it make any sense to prefer living in Charleston in this case, though?
posted by cdward at 10:26 PM on September 29, 2011
Say Charleston had 384,000 murders, and New York had 385,000. NYC stills have more (absolute) crime. Does it make any sense to prefer living in Charleston in this case, though?
posted by cdward at 10:26 PM on September 29, 2011
A lot of people get their image of New York City from gritty 1970s movies like Fort Apache, The Bronx. The truth is that New York's crime rate has been falling steadily since that time. While it's still a big city and there are big city crimes, such as gang violence, muggings, rapes, and the like, it's nowhere near as unnerving a place to live as it must have seemed at the peak of the crime rate.
There are a lot of reasons posited for this, such as William Bratton's use of COMPSTAT to drive policing, or Rudy Giuliani's decision to focus on non-violent crimes such as graffiti that affect the perception of safety by the public, but it's also demographic -- national crime rates have fallen too.
I think the SC guy you encountered just couldn't accept this -- he believe that NYC was a crime-ridden shithole, and so he twisted his logic around the numbers to disprove what they clearly show.
posted by dhartung at 12:49 AM on September 30, 2011
There are a lot of reasons posited for this, such as William Bratton's use of COMPSTAT to drive policing, or Rudy Giuliani's decision to focus on non-violent crimes such as graffiti that affect the perception of safety by the public, but it's also demographic -- national crime rates have fallen too.
I think the SC guy you encountered just couldn't accept this -- he believe that NYC was a crime-ridden shithole, and so he twisted his logic around the numbers to disprove what they clearly show.
posted by dhartung at 12:49 AM on September 30, 2011
(Sorry: meant to add that I lived in NYC for a couple of years in the 1980s, when the scariest years were already in the past, but people still had no idea how much better it was going to get.)
posted by dhartung at 12:50 AM on September 30, 2011
posted by dhartung at 12:50 AM on September 30, 2011
This thread is closed to new comments.
Yeah I'm pretty sure he's just saying that in absolute terms, 10 robberies is 10 robberies and population has no bearing on it
posted by p3on at 3:29 PM on September 29, 2011