Homicides in city jump 50%, other crimes stable - why?

At least not those part 1 crimes which are labeled as "major" crime.

First there is rape, which if you mapped that out, would be a very random set of events all across the city with all sorts of races and backgrounds among victims and offenders.

Then there are car thefts. Car thefts simply aren't what they used to be. I'd say a very small percentage of these crimes are done by gangs who also murder, and another small percentage are those who would steal and joyride a car anytime they saw someone leave their keys in it and run into the store. The majority of them are done by those who have a particular fence for stolen car parts. They are sometimes gang related but more often than not are non violent people.

The majority of burglaries are done by people addicted to crack, heroin or meth who need quick cash. They are not gang bangers, but are probably influenced by some of the drug trade that gang bangers participate in. Other than that you will see a large amount of these burglaries done by a smaller group of people (ie one year one person might commit 200 house breaks, when that person goes to jail, those 200 house breaks simply wont happen, and unless another junkie fills that role, the number might go up or down). Either way, B&Es are usually drug related.

Larceny is so broad that it would be difficult to tie to any sort of murder rate. Since this crime includes shoplifting, it would be interesting to see how many non-shoplifting crimes are involved. Shoplifting is also another crime heavily influcenced by drug use.

Robbery is another crime that is done by drug users, but off the top of my head, I'm trying to think what the percentage of these would be in terms of gang use. The gangbanger is surely going to hold up and rob a gas station, or punch some old man late at night and take his wallet, but there is also a large amount of druggies who snatch bags as well.

I'm not exactly sure what categorizes an "aggravated assault" though, so I'm not so sure how that would reflect on the murder rate. If it includes domestic assaults or drunken street fights, then it probably would have less of an impact on a murder rate than if it simply included only shootings that missed, or other gang related assaults.

And I'm not a stat man myself, but the murder rate is looked at in different terms than a larceny would be since it is so serious. Wouldn't we do the same for a category of "terrorist attacks"?

That's not the difference that's causing all the ruckus. JAK's comment that started the thread relates not to a significant drop from the early nineties to the late nineties but to the significance of the difference between 49 and 73 in 2009 vs 2010. Got any geekitude left for that? Does the blip justify a "Murder up by 50%!" headline, or is it in line with normal random variation for a sample that small?

I'm looking at all the numbers where it's the same every year.

It's generally not appropriate use of statistics to say that a doubling of an insignificant number is suddenly significant. If I'm one gram overweight, and then a few months later I'm 100 grams overweight, I should freak out and go to the nutritionist, right? After all, the amount that I'm overweight increased one hundredfold.

Oh wait, but if you look at the whole picture, that's like going from 140.002 pounds to 140.22 pounds when the chart says I should weigh 140 pounds. Does any of this matter in the least? Of course not. I could gain 1000 grams (a THOUSANDFOLD increase in how overweight I am!) and be 142.21 pounds, and still, any sensible person would round this to saying that I'm still not overweight for any practical purpose.

So you could just as easily round and say that pretty much no one gets homicided in Boston ever.

(And no, I'm not suggesting this, because each life is significant, but just pointing out that it's about as logical as freaking the fuck out that the amount has "doubled" as if we've gone from 20% of our people being killed per year to 40%.)