To Kill, For The Love of Killing

I thought there would be more of a use for this post, but as Mike points out this is contributing to a conversation that isn’t really happening at the moment.  However, I had an interesting time crunching numbers and someone out there might find it useful, so for the morbidly curious, I’m publishing this anyway.

There are essentially three different kinds of multiple murder.

The first is serial murder, when a perpetrator (usually male, usually operating alone) kills multiple people in a series of discrete attacks. The psychology of the serial murder is itself an area of interesting, albeit morbid, study. Famous examples include Jack the Ripper and likely David Berkowitz (the Son of Sam). Serial killers are highly correlated with sociopathic characteristics. One distinguishing factor is that the individual incidents themselves are usually well planned with the intention of allowing the serial killer to commit his or her act of violence and escape to kill again.

The second is political terrorism, when a perpetrator (again usually male, not always acting alone) kills multiple people either in singular or successive attacks. Unlike serial murders, the psychology of the terrorist may or may not be driven by the specific desire for an individual expression of violence. The 9/11 terrorists obviously qualify. The psychology of the terrorist is different from that of the serial killer, as terrorism acts are usually planned and carried out by a number of actors working in concert, thus the group psychology becomes an important factor. The intention of a terrorist event is usually to cause some sort of political reaction. The ability of any one member of the terrorist cell to survive, or to escape capture, is often but not always a secondary factor to the direct outcome itself.

There are difficulties with generalizations in these categories, as certain perpetrators can possibly qualify as one or the other. For example, Ted Kaczynski (the Unabomber) and Timothy McVeigh possesses some overlapping characteristics between terrorists and serial murders. Where they rightfully “belong” is a matter of intentions and his psychology.  Personally, I put Ted in the category of the serial killer and Tim in the next category, as I believe they have more in common with those two containers than with terrorists, but that’s open for debate.

Finally, you have spree killers, when a perpetrator – like serial murders, usually male and usually operating alone – commits essentially a single planned attack, although the single attack can have multiple stages. Unlike serial murders or terrorism, in most cases spree killers end with either the killer committing suicide or “suicide by cop”, and the events themselves are usually staged with the intention to maximize causalities. Unlike terrorism, where the body count is a function of “how terrible” the event can be made, the body count in spree killing events is usually a primary motivational factor.

Contrary to popular opinion, these sorts of attacks are not limited to the United States, nor are they limited to firearms as a primary method of execution. Contrary to political opinion, there is no meaningful correlation between the perpetrators of these events and political leanings; essentially the hold only one thing in common: the perpetrators are violently insane.

From a neuroscience standpoint, this may be an important detail. Charles Whitman, the University of Texas bell tower shooter, was found upon autopsy to have a glioblastoma tumor in his hypothalamus, which may have been pressing on his amygdala, the area of the brain that controls aggression.

Unlike their brethren the serial killers, who still undoubtedly have serious neurological problems, I believe the spree killers likely have a different class of specific neurological disorders that specifically trigger mass aggression. Given the relative paucity of spree killing and serial killing in the general population, it is difficult to make any generalizable claims with a reasonable degree of certainty, however.  Note: that last sentence is probably the most important one in this entire post.

Guns are often used in these attacks, of that there is no doubt.  When intersecting with public policy, then, we must ask the question, “Are guns causally linked to this phenomena, or is there merely a correlation?”  There are two related questions.  The first, “If someone is predisposed to kill large numbers of people, is the relative ease-of-access of guns in the United States a contributing factor in the likelihood of one of these events occurring”, and the second, “If firearms were not available as weapons, would the perpetrators carry out these activities with alternative methods that would have worse outcomes?”

On the 22nd of July, 2011, Anders Behring Breivik used a large car bomb to kill eight people and injure another 209, and then followed up this by using a .223 Ruger Mini-14 semi-automatic carbine and a Glock pistol to kill 69 people and wound another 100, most of them teenagers. It is the record-holding mass killing event using firearms, worldwide.  No fully automatic weapons were used.

On April 27th, 1980, Woo Bum-kon, a South Korean police officer, used two M2 carbines and seven grenades to kill 56 people and injure 35 others before committing suicide. It is the second deadliest mass murder event involving firearms.  Note that his access to weaponry was a function of his employment, and thus any public policy approach to limit this particular event would require disarming the police.

Firearms are not the only weapons, nor are they strictly speaking the most horrific. The largest numbers of victims both killed and injured come from vehicular actions, arson, and bombings.

On September 15th, 2002, Chen Zhengping poisoned a rival shopkeeper’s stock, killing 42 people and injuring several hundred. Rat poison seems to be an increasingly popular weapon of choice in China.

On February 18th, 2003, Kim Dae-han attempted to immolate himself on a crowded subway car. In addition to soaking himself in a flammable liquid, he had two milk cartons filled with some flammable agent. It is unknown what sort of a event he would have carried out, given the chance to follow through of his own accord. Bystanders attempted to stop him from setting himself on fire, and during the struggle one of the containers spilled, eventually resulting in a rapidly-spreading fire that ending 198 lives and injured another 147.  For total injuries and deaths this swamps anything attempted with firearms.

On March 16th, 2001, Jin Ruchao detonated four bombs at different apartment buildings, killing 108 people and injuring 38 more, ostensibly to get back at his former mother-in-law, his ex-wife, and his ex-wife’s new lover. The death toll appears to be ancillary to his desire to blow up the buildings, so this may be an outlier, itself.

In the interest of taking a more rigorous approach, I took the data from Wikipedia’s rampage killer article and encoded it. Some interesting results came out of this exercise.

First, a couple of very important disclaimers: there is huge selection bias in using Wikipedia as a source for this sort of data analysis, for a number of reasons. Wikipedia has an English-speaker population bias for contributors, of course, and this is hardly an unbiased sample. Nevertheless, rampage killings are highly reported in the West, and thus as a preliminary data source it is a much better jumping off point than one particular anecdote. I discarded certain data points from this set. First, I ignored any recorded incident before 1950 as potentially unreliable. Second, I excluded any event that had two cooperative killers, as these would of necessity represent a different psychology than a single killer acting alone (this removes Columbine from the data analysis, 13 deaths and 21 injuries). Any event that would be plausibly regarded as an act of ongoing low-grade war (most of the attacks in Israel, Afghanistan and Iraq during the occupations, the unrest in Chechnya, etc.) were marked and noted where appropriate, as they are not necessarily informative to public policy decision-making for peacetime. I noted where the perpetrator was either active-duty military or law enforcement engaging in mass violence against his own civilian population (side note: in countries in Asia and Africa, these represent the majority of events, not so in the West, although former police and military are not uncommon). Finally, if suicide was directly involved, that was also noted, but only if it could be reasonably inferred that the perpetrator intended his own death (for example, Andrew Stack, who flew his plane into the IRS building in Texas, is labeled a suicide, but the drunk Thai policeman who killed himself and several others with a grenade during a quarrel with his partner is not). Whenever there was ambiguity in the number of victims, I chose the most conservative number; in the rare cases where there was no specificity at all (in all cases this was the number injured, not killed), I encoded it as zero, which will skew the results slightly against total injuries.

Some preliminary observations.

By far the deadliest methodology is suicide by plane, wherein a pilot (or in some cases a passenger) deliberately crashes the plane, immolates it, or detonates an explosive. Of the top 15 deadliest incidents, seven are of this variety. The remainders in this top 15 are arson (four), poison (one), explosives (one), and combination attacks involving firearms and explosives (two).

Attackers who use explosives or vehicles (other than airplanes) are usually not suicidal, whereas attackers who rely heavily on firearms are more likely than any other category to end their rampage in suicide (36 cases of suicide, 54 not, one unreported), even discounting “death by cop”. In the firearms incidents, 66 of the total 93 cases result in the death of the attacker either by suicide, law enforcement shooting, or in a couple of cases vigilante justice.

I did not take an accurate count during the analysis, unfortunately, but the three most popular classes of venue tended toward nightclubs (for explosives or arson), weddings (very popular for the non-firearm related attacks, particularly explosives), and domestic residences (an entire class of the attacks is limited to people who murder their family).

Casualty breakdown

• Firearms-only attacks are responsible for 615 injuries and 767 deaths.
• Firearms together with some other sort of weapon are responsible for 557 injuries and 574 deaths.
• Vehicle-only attacks have 707 injuries and 782 fatalities.
• Poison attacks have 300 injuries and 49 fatalities
• Melee weapons have 64 injuries and 225 fatalities
• Explosives-only attacks are responsible for 1,314 injuries and 266 deaths
• Arson-only attacks are responsible for 470 injuries and 653 deaths

Country comparison
The United States has 33 cases in this data set, the most. China has 18… but with only three exceptions, all of the reported incidents in China take place after 1999, which likely indicates that earlier incidents went unreported in the Western media. Israel has 14, but for obvious reasons their data set is somewhat unique. Germany and the Philippines each have 8 cases, France has 7. Thailand has six (all grenade incidents), as does the United Kingdom. India has five, Indonesia has four. The combined African nations don’t quite meet the U.S.’s total, 21.

If we discount China’s numbers as particularly suspect due to limited reporting, and Israel’s as sufficiently unique to warrant its own analysis, it seems as if the U.S. is a substantial winner over other first world countries, albeit a winner of a dubious honor. However, if we assume that the psychological state that leads to spree killing is distributed randomly throughout a population, we have a more complex picture.

Germany’s 2010 population was about 80 million. The U.S., on the other hand, was 300 million. France was about 60 million, as was the U.K. and Thailand. The Philippines is right about 90 million. Indonesia has a whopping 220 million, India of course tips the scales at over a billion.

Now, the best analysis would be to snapshot each country’s population at the time of the given incidents, but I don’t have that amount of time. Still, as a back-of-the-envelope measure, what we see here is the U.S. having 33 attacks at its population of 300 million, or about 1 incident per 9 million. Germany has one per 10 million, France one per 8.5 million, the U.K. one per ten million… Indonesia, on the other hand, has one per 55 million, and India one per 220 million… compared to the Asian nations, the civilized Western world is significantly more likely to yield a spree killer per capita.  In fact, the difference here is so astonishingly large my immediate impulse is that there is a reporting issue with Indonesia and India, as well as China, and thus the events in those countries that are documented are useful for comparing attack methodologies but likely of little use in establishing comparative frequency of these sorts of events.

Viewing it from this angle, the U.S. isn’t significantly more likely to have one of these events than Germany or France or the U.K. (less than France, actually).

If culture or genetics is a major contributing factor – and they may be – it seems less likely that this is a question of “American culture” vs. “Everybody else” as it is “Western European culture” vs. “Asia and Oceania”.

I should cover Africa in depth, but this post is already getting exceedingly long, and I’m not close to the end yet.

The next relevant question is, how pertinent is the legal status of private gun ownership? Does gun control deter spree killings?

Using the “Gun Rights Index”, found here, we can see the following:

• The United States is 8.0, national (states vary)
• The Philippines is 6.8
• South Africa is 6.0
• Switzerland is 6.0
• Thailand is 4.2
• France is 4.0
• Mexico is 3.9
• India is 3.7
• Germany is 3.2
• Norway is 3.0
• Indonesia is 2.1
• Canada is 2.0
• The United Kingdom at 1.5
• Japan is 1.5
• China is 0.5

Now, an important factor involved here is when various pieces of gun control legislation pass. The United Kingdom, for example, has had several major gun control initiatives pass since the early 1980s, most of them in response to a spree killing event. The efficacy of these is up for debate; a timeline of the U.K.’s six incidents with major pieces of gun legislation is worthy of a post itself, but in a nutshell: 1980 (arson), 1987 (mass killing with two semi-automatic rifles, a pistol, plus arson), 1988 (firearms + explosive), 1988 (1988 Firearms Act, restricting longarms), 1996 (mass killing with pistols), 1997 (Firearms Act of 1997 banned handguns) 1999 (mass killing with explosives), and 2010 (mass killing with firearms).

At a glance, however, I don’t see gun control as a limiting factor on spree killings, in general.

I don’t even see a strong correlation between gun control law severity and method of spree killing. Thailand and the Philippines both have less gun control than most countries, but the preferred method of spree killing in the Philippines is split between melee weapons and explosives, and in Thailand every event is with grenades… in only one incident in either country did the perpetrator use a firearm.

Note, of course, that enforcement is an important issue. The laws on the books in any country may forbid private firearms ownership, but that doesn’t necessarily correlate to ease of acquisition.  This is only a preliminary look at this issue.

I will return to the aforementioned most important line in this post: given the relative paucity of spree killing and serial killing in the general population, it is difficult to make any generalizable claims with a reasonable degree of certainty.  What spree killing means for public policy is thus largely going to be a matter of gut feel and interpretation by those people who argue sides.  The only way to test any claim in this public policy space would be to pass some sort of legislation, and see what happens… but in my own personal opinion, the only thing that will happen is that these events will occur with about the same frequency that they occur now, and the body counts will vary only by the weapons available to the perpetrators.  By the numbers, the average spree killing event using weapons other than firearms is either slightly less (melee weapons being astonishingly good at killing large numbers of people) or much greater.

Now, you may ask what I specifically think, myself.

I think that this sort of violent insanity is probably best dealt with via the mental health system, which is much less robust than I wish it could be.  I suspect that any political capital that is spent to reduce mass murder is best spent increasing access to mental health facilities, which have a number of other awesome public payoffs, rather than attempting to ban or limit gun ownership.  I don’t think there’s much of a case to be made that any real gun control measure other than repealing the 2nd Amendment is legislatively a good use of our representatives’ time as most will run headlong into legal challenges and be overturned, and I don’t think repealing the 2nd Amendment is politically feasible.