Murder by Neighborhood

In the discussion on statistics below, I wondered about trying to recreate the correlation chart by neighborhood instead of by city. I read an article not long ago that posited that the American murder rate is really quite low, outside of certain cities; but, further, that even within those cities the murder rate was quite low outside of particular neighborhoods. It wasn't this article, but this one is the one I can find and it has some good maps.

My hypothesis is that shifting to a neighborhood picture might restore the correlation between violent crime rates and police shootings. I don't know that this is true, but it's the hypothesis that I'd like to test because I think the answer is important. Is there a way to get the relevant data together to try to test that hypothesis?

It is a known issue, of course, that we are dealing with very small numbers and statistical rarity. Perhaps it's not even worth doing, given how little one can really infer from statistics about rare events.

8 comments:

  1. Anonymous11:59 AM

    Would the local university's criminal justice department either have the date or know how to get the date? Maybe one of the police levels, local, state, federal, keep the statistics?

    -stc Michael

    ReplyDelete
  2. I'm guessing they're using the FBI's Uniform Crime Reporting standards. They're the only nationwide ones that I know of; they have some known issues for reliability, but when that's all you have, it's all you have.

    ReplyDelete
  3. Bethany at Graph Paper Diaries is finishing up her master's in statistics and may not have the time, but I'll pass this on and see if she has anything to contribute. She comes over on her own anyway, but not as regularly as I do.

    ReplyDelete
  4. It's true, I don't have time, but I'm a sucker for emails that say "hey go look over at Grim's Hall, there's some interesting data discussions". I downloaded the data set, and a quick glance made me wonder in the opposite direction as Grim...what happens if you pull the state data? A quick and dirty correlation suggests there's a pretty moderate correlation of .48 for state violent crime rates vs state rate of police killing.

    There is zipcode level data in there along with PD, so a neighborhood map might be possible. I'll keep poking.

    ReplyDelete
  5. See, that's the kind of thing that makes blogging worthwhile. It used to happen a lot more often, but it's very nice to see it happen again. :)

    ReplyDelete
  6. Eric Blair8:23 PM

    The Philadelphia Inquirer embarrassed itself in 2009 when it put up an interactive map of shootings in the city, identifying demographically the shooter where known and victim where known.

    Out of something like 2700 reported shootings, something like 2400 involved black people in black neighborhoods.

    The balance was almost all happen in a single neighbor hood that is full of Hispanics.

    They have a monthly 'heat' map of reported crimes and all the violent crimes are pretty much were one would expect them.

    Solve for demographics, and the USA if a very safe country.

    ReplyDelete
  7. Gringo10:49 PM

    Solve for demographics, and the USA if a very safe country.
    Pretty much so.Not that such a demographic analysis is any comfort for blacks.
    From 538.com:
    Black Americans Are Killed At 12 Times The Rate Of People In Other Developed Countries.

    Homicide Deaths per 100,000 Population
    2010-2012 average, High Human Development Countries


    United States, black 19.4
    United States, Hispanic 5.3
    United States, overall 5.2
    United States, white 2.5
    Finland, 2.0
    Israel, 1.9
    United Kingdom, 1.1
    Netherlands, 0.9

    Whether the U.S.’s homicide rate qualifies as an outlier depends on exactly where you set the cutoff for an “advanced” country. Among countries with an HDI of .850 or higher — these are the 31 most well-off countries in the world — the U.S.’s rate of homicide deaths, 5.2 per 100,000 persons, is easily the highest. The next-highest are Brunei (2.0), Finland (2.0) and Israel (1.9). And the U.S. homicide death rate is more than three times higher than neighboring Canada (1.5).
    The U.N., however, sets a slightly lower threshold for a developed country, describing all countries with an HDI of .800 or higher as having “very high human development.” Several countries with an HDI between .800 and .850 have a homicide death rate comparable to the U.S., including Lithuania (6.9), Argentina (5.5), Estonia (5.2), Cuba (4.2) and Latvia (3.8). The U.S.’s homicide death rate is also much lower than Mexico’s (22.0), though Mexico’s HDI is just .755.
    But these comparisons neglect a massively important fact, and one that is especially pertinent in the wake of the Charleston shootings.
    Extending on an analysis by the academic Kieran Healy, I calculated the rate of U.S. homicide deaths by racial group, based on the CDC WONDER data.3 From 2010 through 2012, the annual rate of homicide deaths among non-Hispanic white Americans was 2.5 per 100,000 persons, meaning that about one in every 40,000 white Americans is a homicide victim each year. By comparison, the rate of homicide deaths among non-Hispanic black Americans is 19.4 per 100,000 persons, or about 1 in 5,000 people per year.
    Black Americans are almost eight times as likely as white ones to be homicide victims, in other words.
    So for white Americans, the homicide death rate is not so much of an outlier. It’s only modestly higher than in Finland, Belgium or Greece, for instance, and lower than in Chile or Latvia.

    ReplyDelete
  8. BTW - for those of you not familiar with Chicago, that blank space in the upper left is not a mistake by the map maker. That space is occupied by Harwood Heights and Norridge, two separate municipalities that are independent of Chicago but completely surrounded by it. Collectively they are about 95% white and > 1% black, with the rest being mostly Asian and a significant amount of Native American. About 4% of the population is Hispanic of any race. Lots of people with Polish or Italian ancestry.

    ReplyDelete