4 ms·

The result I've heard about may just be localized, for example in New Orleans: http://www.nola.com/opinions/index.ssf/2014/07/new_orleans_murder_rate_drop_i.htm

The result I've heard about may just be localized, for example in New Orleans: http://www.nola.com/opinions/index.ssf/2014/07/new_orleans_murder_rate_drop_i.html http://www.nola.com/opinions/index.ssf/2014/07/new_orleans_m... (there are several similar stories for other crime-prone metropolitan areas). Simpson's Paradox can also mess with these statistics: http://en.wikipedia.org/wiki/Simpson%27s_paradox http://en.wikipedia.org/wiki/Simpson%27s_paradox – like, is crime decreasing because we are suburbanizing? And therefore, if we hold the community type constant, is crime decreasing? And so if you ask someone "is crime decreasing in your area of perception?" then someone can reasonably answer "no". (There's also the possibility of a fallacy where everyone ratchets up their expectation – e.g., a person who moves to the city from the suburbs says crime increases because it probably has for them, but a person who moves to the suburbs from the city starts looking to media to make observations about their former neighborhood.)

mikeyouse 12y ago
The math doesn't add up, and that's not data being presented from the article about New Orleans, it's an opinion piece. In 2009, there were 424 non-fatal shooting victims and 174 homicides in New Orleans. In 2013, there were 321 non-fatal shootings and 155 homicides.[1] The population of New Orleans was 354,850 in 2009 and was 378,715 in 2013.[2] So the total murder rate went from 49/100k to 41/100k (17% decrease) while the total non-fatal victimization rate went from 120/100k to 85/100k (30% decrease). To see whether or not better care made an impact, we can just multiply the number of injured by the difference in survival rates after the advanced trauma methods were introduced (14% vs. 19%). The number of non-fatal shootings would have decreased by 16, and the number of murders would have increased by the same amount. So the new rates would be 45/100k and 81/100k, still 8% and 33% improvements in the past 5 years. Simpson's Paradox doesn't cut it either; young vs. old, rural vs. urban, big city vs. little city vs. small town, black vs. white vs. hispanic, West vs. Midwest vs. South --- All rates are down across the board.[3] [1] - http://www.nola.com/crime/index.ssf/2014/01/2013_new_orleans_murder_rate_d.html http://www.nola.com/crime/index.ssf/2014/01/2013_new_orleans... [2] - http://www.census.gov/popest/data/counties/totals/2013/index.html http://www.census.gov/popest/data/counties/totals/2013/index... [3] - http://www.bjs.gov/content/pub/pdf/fv9311.pdf http://www.bjs.gov/content/pub/pdf/fv9311.pdf
- ianbicking 12y ago
  I'm not deeply wedded to this particular point about trauma care, I think increased risk adversity and security measures are more important to the perception of crime (and give a rational basis to a perception of the threat of crime). But this article is at least a bit more direct, and refers to the things I've read in the past (and some of those articles I admit probably contained false assertions): http://www.separatinghyperplanes.com/2013/07/can-medical-advances-explain-falling.html http://www.separatinghyperplanes.com/2013/07/can-medical-adv... Per that article, the data supporting better trauma care is not particularly impressive, but it does seem like there is a discrepancy between the criminal reporting of assault with a firearm, and the medical reporting of injury from a firearm

mikeyouse 12y ago

The math doesn't add up, and that's not data being presented from the article about New Orleans, it's an opinion piece. In 2009, there were 424 non-fatal shooting victims and 174 homicides in New Orleans. In 2013, there were 321 non-fatal shootings and 155 homicides.[1] The population of New Orleans was 354,850 in 2009 and was 378,715 in 2013.[2] So the total murder rate went from 49/100k to 41/100k (17% decrease) while the total non-fatal victimization rate went from 120/100k to 85/100k (30% decrease). To see whether or not better care made an impact, we can just multiply the number of injured by the difference in survival rates after the advanced trauma methods were introduced (14% vs. 19%). The number of non-fatal shootings would have decreased by 16, and the number of murders would have increased by the same amount. So the new rates would be 45/100k and 81/100k, still 8% and 33% improvements in the past 5 years. Simpson's Paradox doesn't cut it either; young vs. old, rural vs. urban, big city vs. little city vs. small town, black vs. white vs. hispanic, West vs. Midwest vs. South --- All rates are down across the board.[3] [1] - http://www.nola.com/crime/index.ssf/2014/01/2013_new_orleans_murder_rate_d.html http://www.nola.com/crime/index.ssf/2014/01/2013_new_orleans... [2] - http://www.census.gov/popest/data/counties/totals/2013/index.html http://www.census.gov/popest/data/counties/totals/2013/index... [3] - http://www.bjs.gov/content/pub/pdf/fv9311.pdf http://www.bjs.gov/content/pub/pdf/fv9311.pdf

ianbicking 12y ago
I'm not deeply wedded to this particular point about trauma care, I think increased risk adversity and security measures are more important to the perception of crime (and give a rational basis to a perception of the threat of crime). But this article is at least a bit more direct, and refers to the things I've read in the past (and some of those articles I admit probably contained false assertions): http://www.separatinghyperplanes.com/2013/07/can-medical-advances-explain-falling.html http://www.separatinghyperplanes.com/2013/07/can-medical-adv... Per that article, the data supporting better trauma care is not particularly impressive, but it does seem like there is a discrepancy between the criminal reporting of assault with a firearm, and the medical reporting of injury from a firearm