Alright, let’s talk homicide by gun. If we look at the general statistics (page 87 here: https://www.cdc.gov/nchs/data/nvsr/nvsr65/nvsr65_04.pdf), we find that a total of 15,872 people were killed by others (homicide). Of these, 11,008 (or 69%) were killed by firearm. The next known cause of death is cut/pierce. Note, though, that there are a large number of homicides (1,585) where the cause of death is unspecified, and there are 4,597 deaths where the type of death is undetermined (it could very well be that some of those are homicides, which would change the numbers somewhat). Nevertheless, let us proceed with the cases where we know the type and cause of death.
Let’s look at the breakdown in terms of relationship between killer and victim in cases of homicide (https://ucr.fbi.gov/crime-in-the-u.s/2014/crime-in-the-u.s.-2014/tables/expanded-homicide-data/expanded_homicide_data_table_10_murder_circumstances_by_relationship_2014.xls). Some interesting notes here:
1. The single biggest (known) circumstance category in terms of total murder victims is “Other arguments” at 2,786. Within this, the biggest relationship is Acquaintance.
2. “Gangland killings” (defined as a murder carried out by organized criminals (https://en.wikipedia.org/wiki/Gangland_killing)), adult and juvenile, total 715–6% of all murders as tracked by the FBI.
So these are interesting, right. Because the biggest (known) circumstance surrounding homicide is a type of argument between acquaintances (which apparently includes homosexual couples — this is from 2014). Further, organized criminals account for a relatively small percentage of total murders. Note that this also doesn’t include drug traffickers (the category labeled “Narcotic drug laws”) which comes under “Felony type” rather than “Other than felony type”. This would suggest that killings between gangs (or by gangs) account for a relatively small proportion of the total murders. Just a note that the circumstances of the murder are unknown for 37.7% of the murders tracked here, and I’m not sure how those would spread out — for example, if a majority of those happened to fit in one main category or sub-category, that would skew the numbers with respect to circumstance. In the same vein, the relationship of killer with victim is unknown for 45.5% of murders, so it’s also unclear if the statistics are skewed by that. I am also unclear as to why the FBI recorded a different number of homicides (11,961) than the CDC (15,872) — maybe they are using different definitions?
Now let us dive into statistics based on race and sex (page 67 here: https://www.cdc.gov/nchs/data/nvsr/nvsr65/nvsr65_04.pdf). Some things stand out:
1. Black males are the biggest risk group for dying by homicide, with Native American males being the next biggest risk group.
2. However, from above, it’s unclear whether most of this is due to gang activity (which is what some people claim).
So we know that the tool used most often in homicide is a gun, black males are most likely to be killed, and gang violence most likely accounts for a small percentage of total homicides. So now the question is: what do we do?
Honestly, I am not sure. If many of these murders are spontaneous (which seems plausible given the distribution of circumstances), it seems reasonable that restricting gun ownership wouldn’t really do much. Why? Because in the heat of the moment, they will pick up some other weapon or improvise one out of other stuff around them — this wasn’t a planned murder and their objective is to hurt/kill the other person (for whatever reason). And while there *may* be a slight uptick in survival after the attack if a gun is not used (for example, if someone tries to use a knife and is just really bad at it or stabs in the wrong place, or if they aren’t strong enough to hit someone over the head with some blunt object), I’m not sure how much that effect would be — it’s certainly something to look into. That is, if the survival after attacks would go up an appreciable amount (as in the case of suicide), there is a case to be made that banning all guns would help decrease (successful) homicides. However, that is a big if that needs to be studied in detail.
Looking at a couple of different datasets here (https://webappa.cdc.gov/sasweb/ncipc/mortrate.html and https://webappa.cdc.gov/sasweb/ncipc/nfirates.html), we get an interesting picture (I am looking at the Homicide rates in the first one and Assault rates in the second one):
1. 1,740/131,090 = 1.3% of death by cut/pierce
2. 56/10,496 = 0.5% chance of death by transportation (all)
3. 11,008/71,478 = 15.4% chance of death by gun
4. 89/6,433 = 1.4% chance of death by fire
In each of these cases, I looked at deaths / (deaths + nonfatal injuries), which should give a rough estimate of rate of death from each of these tools. Of course, one assumption I am making here is that the cases of assault were meant to be fatal and failed, which is a large assumption. If this is the case, though, it seems that guns *are* much more fatal than than other methods of assault and murder. That is, it could very well be that banning all civilian gun ownership would help bring down successful homicide rates.