Saturday, November 07, 2015

Black cities and dead bodies

Steve Sailer, commenting on the how the head of the DEA, following similar remarks from the head of the FBI, recently insinuated that the year-over-year increase in violent crime is in part a consequence of police being reluctant to confront criminal behavior for fear of their every move being subject to intense (and hostile) public scrutiny:
One way to check this is to look at my graph of the 60 biggest cities ranked in order of increase in murder victims from pre August 9, 2014 to the same time period in 2015 (based on Carl Bialik’s data in FiveThirtyEight.) The biggest absolute increases have been in heavily black cities.

The Obama Administration’s promotion of BlackLivesMatter rageaholics seems to be encouraging blacks to murder blacks.

Okay, heavily black Baltimore tops the chart, but Memphis--which has a population that is almost identically black (63.1% to 63.3%, respectively)--is second to the bottom. Chicago is near the top, but Raleigh, a city in the South, has seen an absolute decrease in homicides. It's basically a wash. Racists will see what they want to see!

Fortunately, regression analysis can help resolve this dispute. Using Census quick facts data, we find that the correlation between the percentage of a city's population that was black in the 2010 census and the absolute change in the number of homicides it experienced in the first seven months of 2015 compared to the first seven months of 2014 is .44 (p-value = .0006).

When a single variable correlates this robustly with an outcome in the social sciences, it's worth paying attention to, even more so when we take into consideration potential discrepancies in the geography of police jurisdiction in 2015 and the geography of Census city designations in 2010. Noise and imprecision in measurement will virtually always weaken perceived correlative relationships. In other words, the true relationship is probably even higher than .44.

Steve is one of the best in the business at noticing patterns, so it comes as little surprise that his general impression has empirical validation.

A plausible and seemingly obvious consequence of the "black = good, white = bad" narrative, embodied most saliently in the Black Lives Matter movement, is that blacks feel let off the leash. A corollary to that, though, is that whites tend to feel even more restrained (at least until they come to a breaking point and abruptly and savagely tear the arm off of the hand that has been slapping their cheeks for a generation).

When blacks, who already behave relatively poorly, are encouraged to behave with still less restraint than before and relatively well-behaved whites are simultaneously shamed for even allowing a putatively improper thought to materialize in the recesses of their minds, the gap in civilized behaviors between blacks and whites grows yet wider. Falsely portraying predatory blacks as victims pushes blacks to be more predatory. Falsely portraying white victims as predators pushes whites to become more sheepish (again, for some period of time before the threshold of tolerance is breached).

For awhile, it's great for race-hustling grievance mongers and virtue-signalling ethnomasochistic whites alike. Good things have a tendency of coming to an end, though.

15 comments:

enLightened said...

Correlation is not causation
Repeat after me
Correlation is not causation
Correlation is not causation
Correlation is not causation


Anonymous said...

enlightened: you are, of course, free to say that, but it will be entirely unpersuasive unless you can name some other plausible factor which has simultaneously increased primarily in increasingly black cities which directly causes crime - which is *not* related to, you know, the worst racial political crisis in a generation which happens to center around policing and crime...

Noah Carl said...

Nice post. Do you have a scatterplot?

TangoMan said...

AE, you'll probably find the Siegel et al paper to be of interest in that they looked at predictors of homicide. Here's the paper. Click the links for tables 2 and 3 to see the strength of various predictors of firearm homicide.

If you've clicked the links and seen the news, then pop over to ThinkProgress and read their report on this paper and notice their laser-like focus on firearm ownership as a predictor of firearm homicide. They found this piece of information newsworthy:

A one standard deviation change in firearm ownership shifted gun murders by a staggering 12.9 percent.

But didn't think that this finding merited any mention:

For each 1-SD increase in proportion of black population, firearm homicide rate increased by 82.8%

If a 12.9% increase is staggering, then how should we characterize an 82.8% increase?

Anonymous said...

I don't disagree with what you said in this post. I will however, say that both the extreme left wing and the extreme right wing tend to focus on negative stuff so much. Every time some study comes out that states "black children are falling behind in academics" left wingnuts shriek and clamor for solutions to "close the gap". Every time some other study comes out that states "blacks have a higher rate of violence than other peoples" right wingnuts shriek and clamor for genocide or forced deportation.

I'm not going to do either of these things. I'm going to look at the happy news. Boston has the largest DECREASE in homicides! My guesses as to why this happened:

1. increase in low crime demographics (likely)
2. decrease in high crime demographics (unlikely)
3. all/most demographic groups have better moral values than they used to (unlikely)
4. rate of attempted homicides stayed the same; quality and quantity of doctors/surgeons went up (most likely)

In short, I believe that Boston has the largest decrease in homicides because Boston is the kind of locale that attracts doctors and med students.

Audacious Epigone said...

Enlightened,

It's not necessarily causation, but it might be. Or it might be part of the causal explanation. It provides more value than ungrounded assertions do.

Noah,

I don't see contact information. I'll email it to you. Creating data labels on scatter plots is tedious with my old version of excel.

Tangoman,

Don't ask a question like that, just refer to this.

That's really intellectually dishonest. It's not as though it can even be passed off as an instance of inadvertently overlooking relevant data. The black effect is right below the firearm ownership effect in table 3.

Anon,

This is only over the course of a year. How much have those four variables changed over that short period of time?

Noah Carl said...

Audacious Epigone, my email is noah.carl@nuffield.ox.ac.uk.

Dan said...

The data is actually incredibly robust, because each killing represents several people being shot, since most people who get shot are saved. So Baltimore's 300+ homicides actually represents upwards of a thousand attempted murders that score a hit, and those are a subset of all attempted murders including those that miss completely.

Furthermore, in Baltimore, the number of homicides for the past 6 months is equal to the number in all of 2014, so you really have a doubling of the homicide rate.

And more, the rate has actually been increasing, with 10 homicides in the last week including a cold and rainy weekend that kept people indoors.

http://data.baltimoresun.com/bing-maps/homicides/index.php?show_results=UPDATE+MAP&range=7&district=all&zipcode=All&cause=all&age=all&gender=all&race=all&article=all

Maybe the rate will drop on a day in January if there is a really immobilizing snowstorm.

Obama is in a bind. He has to keep moving leftward in order to stay cool. If he is not shouting "F* tha police!" soon, he will fall way behind the cool kids.

Dan said...

More on Baltimore, non-fatal shootings are actually up much more sharply than homicides, meaning that those ER docs have actually upped their game considerably.

JayMan said...

"we find that the correlation between the percentage of a city's population that was black in the 2010 census and the absolute change in the number of homicides it experienced in the first seven months of 2015 compared to the first seven months of 2014 is .44 (p-value = .0006)."

What about change in the homicide rate?

Joshua Sinistar said...

This one's easy. The Po-leese are stepping aside and letting the apes kill each other. Its just a job for them. Put in twenty and retire. No one wants to go to Nigger University for nailing down a pavement ape.
I'm not sure exactly what the Eye of Soros and his BLM movement are trying to do. Why the Government is tossing cops to the wolves when people are talking Revolt is stupid beyond belief, but then we're not talking competence with these people.
I suppose they're hoping to get more joke black and meskin cops who will shoot at Whites. Go ahead fuckers. Those are watercolor people. All they do is run. Ever see a whole group of Urban Wildlife scatter when just one cop comes rolling by? Courage is not found in the Urban Jungle.

Audacious Epigone said...

Jayman,

That'd take more time to figure out. Calculating homicide rates over a seven month period for 60 cities is a lot of additional tedium for a result that, from eyeballing the data, doesn't look like it'll change the thrust of the outcome much.

Dan,

So if I ever find myself stranded in Baltimore, make my way to the north center of the city!

Joshua,

The term I've started hearing bantered around is "depolicing". It's exactly what it sounds like it would be. Want to live in a world of sheep and wolves with no sheepdogs? Enjoy.

Anonymous said...

"I'm not sure exactly what the Eye of Soros and his BLM movement are trying to do."

Kill black people, apparently. It is like abortion which also disproportionally kills blacks in the name of "saving" them from poverty. BLM (ironically the same acronym as the Bureau of Land Management) is all the rage on college campuses as a mark of righteousness for the current crop of junior do-gooders. BLM has not helped blacks. Heck look how many more have died. Have there been fewer fatal police shootings? If so, how many fewer than last year? Because if the homicide uptick in Baltimore and St. Louis is greater than the number of fewer killed by police in those cities, I think the BLM is mighty Orwellian or it really is a program of the shadow Bureau of Land Management.
http://www.defenders.org/press-release/defenders-wildlife-challenge-blm%E2%80%99s-first-ever-approval-wolf-hunting-derby-public-lands

Dan said...

"So if I ever find myself stranded in Baltimore, make my way to the north center of the city!"

Yes! That would be the Johns Hopkins campus, the Loyola College campus and campus of the College of Notre Dame. It is quite interesting how perfectly other things sort themselves out, once you select for IQ.

Audacious Epigone said...

Ha, I wonder if Steve Sailer has realized that BLM is an acronym shared by black lives matter and the bureau of land management.