Suicide, Homicide and Guns

There are a number of reasons social psychologists are confident that the availability of firearms is a major cause of violent death. In an earlier post, I discussed some of the many studies of the priming effect of weapons on aggression. These studies suggest that the presence of weapons, or pictures of weapons, increases aggressive thoughts and behaviors. Since these are laboratory experiments in which participants are randomly assigned to conditions, they permit a relatively strong inference of causality. However, due to the obvious constraints on our willingness to provoke aggression in the lab, some people find these studies to be artificial, and feel more comfortable with field studies of the availability of firearms in the real world.

Many of these field studies are ecological, in the sense that they measure both the availability of weapons and the number of deaths at the aggregate or population level. For example, in an early study, Archer and Gartner (1987) found that the homicide rate in countries around the world is highly correlated with the availability of handguns. But correlation does not mean causation, and the countries involved in this comparison differed in many other ways besides their weapons policy. Most recent studies have attempted to statistically control for alternative explanations such as urbanization, poverty, alcohol use, depression, etc. However, since the number of alternative explanations is theoretically infinite, there are limits to this approach.

© nypost.com

Gun advocates also attack the validity of the measures used in these studies. For example, in most studies, the availability of firearms is determined from gun registration statistics. However, it could be argued that many homicides are the result of unregistered guns in the hands of criminals. Can this assumption account for the positive correlation between gun availaility and murder? Yes, but only if you assume that those areas that contain the most legally registered guns also contain the most illegal ones.

A more time-consuming but more conservative procedure is to start with a suicide or homicide and work backward, looking at each individual case. In these observational studies, for each suicide or homicide victim, regardless of the manner of death, it is determined whether firearms were available in the home. These people are compared to a matched control group of people who did not commit suicide or were not murdered, such as participants in a national health survey. This allows you to compute a likelihood ratio—the likelihood of a homicide or suicide when a gun is available compared to the likelihood when it is not.

In a study published this week, Andrew Anglemyer of the University of California at San Francisco and his colleagues did a thorough literature search and located 15 such observational studies. They then did a meta-analysis of these data. A meta-analysis is an analysis of analyses, a way of statistically combining the results of several studies to estimate the size of an effect. These were their findings:

• You are 3.24 times more likely to commit suicide when a firearm is readily available than when it is not. These findings do not differ for men and women.

• You are 1.94 times more likely to die of a homicide with a gun in the home than without. This effect differs for male and female homicide victims. Women are 2.94 times more likely to be killed when a firearm is available, while for men, the effect is almost nil—only 1.29 times more likely.

The meta-analysis shows that having a firearm available has a greater influence on suicides than on homicides. This makes sense, because most firearm suicides kill themselves with their own gun, while most firearm homicide victims are killed with someone else's gun.

The homicide data show that firearm availability only increases the risk for female victims. Other studies show that women are most likely to be murdered by someone they know. This suggests that firearm availability makes it easier for domestic violence to become lethal. On the other hand, there is no evidence that owning a firearm increases the risk of homicide for men, since they are usually not killed with a gun belonging to themselves or a family member.

In an accompanying editorial, David Hemenway, a leading authority on firearm victimization, points out that this pattern of results is due to Anglemyer's choice to analyze only individual-level, observational studies. He argues that the meta-analysis substantially underestimates the effects of firearm availability on male homicides because it only asks whether the victim had access to a gun, when the more important question is whether the perpetrator had access to a gun. Therefore, he suggests that future studies of homicides be either population-level ecological studies which measure the number of guns in the community, or individual-level observational studies of the availability of firearms to perpetrators rather than victims.

You may also be interested in reading:

Symbolic Racism and Gun Ownership

Open Carry

Breathing While Black

Situation Alarming--But Not Serious

It's sometimes said that liberals are concerned about violence in the mass media but not sexuality, while conservatives are concerned about sexuality but not violence. If so, then a new study by social psychologist Brad Bushman and his colleagues suggests that the Motion Picture Association of America's (MPAA) rating system is a system of conservatives, by conservatives, and for conservatives.

The PG-13 rating was introduced in 1984 to solve a financial problem for the film industry: Too many popular films were being rated R, and an R rating meant that children under 17 could not attend unless accompanied by a parent or guardian. According to the MPAA, a PG-13 film “may go beyong the PG rating in theme, violence, nudity, sensuality, language, adult activities or other elements, but does not reach the restricted R category.” With respect to violence, they say, “There may be depictions of violence in a PG-13 movie, but generally not both realistic and extreme or persistent violence.” So, how's it going?

First, the methodology. Bushman drew up a list of the 30 top-grossing films of each year from 1950 to 2012 (1890 films), from which they randomly selected half, or 945 films. Trained coders watched the films and identified violent sequences, defined as “physical acts where the aggressor makes or attempts to make some physical contact with the intention of causing injury or death.” This definition excludes natural disasters, accidents and sports injuries. A sequence refers to the continuous action of a character, regardless of the number of discrete acts in the sequence. This solves an important practical problem, because it is not necessary to count the exact number of violent acts, i.e., punches, gunshots, which is a major source of unreliability in measuring violence. Using this approach, they achieved a reliability of α = .80, which is pretty good.

(I have very few questions about this approach, but one problem that concerns me is the scoring of scenes with large numbers of participants, such as battle scenes. Since there is one violent sequence for each character, how do you measure the number of participants? And how do you prevent such scenes from skewing the data?)

Next, they divided the films into 5-minute segments. The number of violent sequences in each segment was counted, and each film was given a score representing the rate of violent sequences per hour. In case you're wondering, 94% of all films between 1950 and 2012 had at least one violent sequence.

Bushman also counted the number of sequences involving guns, a gun being defined as “a weapon that can be carried with one or both hands that fires a bullet or energy beam with the intention of harming or killing a living target.” Hunting and target shooting were excluded. The reliability in scoring these sequences was even better (α = .91). A violent scene involving a weapon is less ambiguous than other forms of violence.

For purposes of analysis, films after 1984 were divided by rating into three categories: G and PG, PG-13 and R. G and PG were combined because there were fewer films with these ratings among the box office hits.

First, Bushman looked at the overall trend from 1950 to 2012. Not surprisingly, there was an increase. The level of violence more than doubled. The rate of increase was not simply linear, but gradually accelerated over the 63-year period.

Then they looked at gun violence during the period from 1984 to the present. Again, there was an increase, but there were important differences by rating. The G- and PG-rated films averaged 1.26 gun sequences per hour, and the overall rate declined slightly. The R-rated films averaged 2.15 sequences per hour and the rate did not change. However, violence in PG-13 films increased significantly, more than tripling over the 29 years. It started at near-zero, but accelerated sharply in recent years. From 2009-2012 the rate of gun violence was higher in PG-13-rated films than in R-rated films, flatly contradicting the MPAA guidelines.

The study does not report any measures of nudity or sexuality in films, but one of the co-authors, Daniel Romer, said that they measured sexuality, and it is much more likely to get an R rating than violence. I presume these results will appear in a future publication.

The relationship between filmed and televised violence and aggressive behavior is one of the strongest findings in social science, replicated in hundreds of experimental and correlational studies. The American Psychological Association and several other professional organizations have issued position papers highlighting these findings and recommending public policy changes. When you statistically combine the results of these studies in a meta-analysis, the relationship between filmed violence and aggressive behavior turns out to be almost as strong as that between smoking and lung cancer, and higher than many other statistical relationships that society takes very seriously, i.e., exposure to lead and IQ, homework and academic achievement. Exposure to media violence also makes participants less likely to help a person in need.

Daniel Craig as the PG-13-rated James Bond

Gun violence is a particular concern due to the weapons effect. Experiments show that the mere presence of a weapon—or a picture of a weapon—in the environment increases aggressive behavior in both angry and non-angry participants, which is probably why the presence of a gun in the home increases homicides.

While preparing this blog entry, I read at least a dozen reports about Bushman's study in the corporate media. Sadly, most of them contained false balancing (such as this one and this one). That is, they “balanced” Bushman's findings with comments from skeptics who dismissed the idea that violence in films might influence real-world aggressive behavior. (Science Daily was a notable exception.) The most common argument was that homicides have declined in this country in recent years, even as filmed violence has increased. This reasoning is, of course, completely fallacious, since there are many other variables that influence the homicide rate.

False balancing creates the mistaken impression that “experts disagree” over whether gun violence in movies really matters. Considering the popularity of violent films and television shows, this may be exactly what their audience wants to hear. Bushman has recently written an article suggesting several ways people rationalize away these findings. It is also worth noting that most of these news organizations are owned by larger media corporations that profit from the sale of filmed and televised violence. Bushman and Anderson have found that as the findings linking media violence and agression have gotten stronger, the news media have become more assertive in dismissing them.

You may also be interested in reading:

Open Carry

Breathing While Black

The Job Killers, Part 1

Open Carry

Last week, in connection of the shooting death of Trayvon Martin by George Zimmerman, I posted research showing that people are more likely to “see” a gun in the hands of a African-American man than a white man, even when that black man is carrying a harmless object such as a cell phone. Now I've found a new series of studies that suggest that carrying a gun also increases the likelihood that you will see a similar weapon in the hands of others.

Photo by itwuzcryptic

In a soon-to-be-published paper, Jessica Witt and James Brockmole of Notre Dame University report five studies in which they asked student participants to look at slides of men carrying either guns or harmless objects. The students had about a half-second to indicate whether the person was armed or not by raising or lowering a lever. The experiment varied the handle of that lever. In one condition, it was shaped like a realistic-looking toy handgun. It the other, it was a rubber ball. The results showed that the participants were more likely to “see” the person in the slide as holding a gun when they themselves were holding a gun than when they were holding a ball.

The authors interpret these findings as consistent with the theory of event coding. This theory is a scientific variation on the old saying that when you're holding a hammer, everything looks like a nail. When planning an action with an object or tool, you are predisposed to identify other objects on which the tool might be used.

The authors also considered the theory of conceptual priming. This theory suggests that the mere presence of a weapon in the environment predisposes people to “see” other weapons. They rule this explanation out due to one of their five experiments. In this variation, a real firearm was placed on the table while the subjects were performing the task using the lever with the rubber ball handle. There was no greater tendency to see the man in the slide as holding a firearm when a real gun was present in their environment than when it was not.

I suspect that they ruled out the priming explanation prematurely. First of all, there are other studies showing that aggressive cues in the environment affect behavior. For example, in a study by Berkowitz and LePage, frustrated participants were more likely to respond aggressively when there were weapons (a handgun and a rifle) on the table than when there was either a neutral object (a badminton racquet) or nothing on the table. Secondly, in the Witt and Brockmole study, participants were exposed to many photos of handguns, and guns were referred to when explaining the task. It's possible that in such a gun-saturated environment, the presence of yet another gun—albeit a real one—didn't make much difference.

I go into these details not only because I believe in priming—an effect supported by literally thousands of studies. It also affects the practical implications of Witt and Brockmole studies. A strict interpretation of them would state that the bias to “see” a gun occurs only when you are holding a gun in your hand.

In this country, all but eight states permit ordinary people to openly carry a firearm either on their person or in their car (although some of them require a license), and all but Illinois (and D. C.) allow people to carry concealed weapons. This research implies that carrying a firearm makes you more likely to see others as dangerous. Does it matter whether you are carrying the gun in your hand, in a holster, or in the glove compartment of your car? Probably. You are more likely to be aware of the gun when it's in your hand (as Zimmerman's apparently was). But I suspect that carrying a concealed weapon alters threat perception as well. Maybe future studies will test the limits of this effect.

Breathing While Black

Beginning ten years ago, social psychologists did a series of studies looking at the consequences of our cultural stereotypes about African-American men and violence. A study by Joshua Corell and others called “The Police Officer's Dilemma” is typical. White participants were shown slides of young men standing in public places, such as a park or a city sidewalk. Half of the pictures were of black men and half of white men. Within each racial group, half of the men were holding handguns and the others were holding some innocuous object, such as a cell phone or a soda can. The participants had a half-second to press one of two keys labeled “shoot” or “don't shoot.” The researchers counted the number of errors—either not shooting an armed man or shooting an unarmed man.

The results showed evidence of bias. When the men in the photos were white, the participants made the same number of errors regardless of whether the men were armed or not. However, when the men were black, they made fewer errors when they were holding a gun, and more errors when they were not. In other words, they were more likely to “shoot” a black man than a white man, both when he was armed and when he was not.

This is an example of automatic or fast thinking, which Kahneman calls “a machine for jumping to conclusions.” Subsequent research showed that the results did not vary with personal prejudice. However, the participants showed a greater “shooter bias” when they were more aware of the cultural stereotype, that is, when they gave higher estimates of the percentage of Americans who saw black men as aggressive, dangerous and violent. Exposure to newspaper articles about black criminals also increases the size of the shooter bias.

The fact that this response tendency is automatic doesn't mean it can't be controlled. The shooter bias can be reduced through practice and with certain kinds of training, such as instructing participants to ignore the race of the person in the photo and to concentrate on the presence or absence of a gun.

This series of three videos (about 35 minutes total) featuring social psychologist Jennifer Eberhardt deals with primarily unconscious racial biases. The study on weapons stereotypes is in the second segment, but I recommend them all. (You Tube will guide you through them.) Some of them are shocking, and will make it clear why social psychologists are unwilling to accept mass media assurances that prejudice is a thing of the past.

I think these studies are relevant to the killing of unarmed black teenager Trayvon Martin by Florida neighborhood watch volunteer George Zimmerman. Zimmerman says he felt threatened by Martin, and so far, the Sanford, FL police have accepted his claim that he acted in self defense. This defense is made possible by Florida's 2005 “stand your ground” law, which was written by the National Rifle Association. Variants of which have been enacted in 20 other states. The relevant section of the law reads as follows:

A person who is not engaged in an unlawful activity and who is attacked in any other place he or she has a right to be has no duty to retreat and has the right to stand his or her ground and meet force with force, including deadly force, if he or she reasonably believes it is necessary to do so to prevent death or great bodily harm to himself or herself or another or to prevent the commission of a forcible felony.

This law does away with the longstanding legal doctrine that, when outside your home, there is a “duty to retreat” when confronted with a dangerous situation. It creates ambiguity about when a person is acting in self-defense. How can it be objectively determined that a person “reasonably believes” he or she is threatened? The studies of shooter bias suggest that people might sincerely believe themselves to be in danger when confronted with an unarmed black man. Even if the perpetrator is wrong, he or she may escape punishment if the mistake is “reasonable.”

The difficulty is compounded when there are no witnesses or the witnesses are friendly to the perpetrator. Who is able to contradict the shooter's self-report of his or her emotional state? Not the victim. He's dead. As a result, police and prosecutors may assume that they have little chance of obtaining a conviction in a jury trial.

From reading the accounts or this incident and listening to the 911 tape (see below), I doubt whether Zimmerman sincerely felt himself to be threatened. He pursued Martin even after the police told him not to. However, the reference in the law to “prevent(ing) the commission of a forcible felony” seems to encourage this type of vigilantism. When combined with Florida's lax gun laws, the “stand your ground” law allows armed, aggressive, and possibly paranoid people to pursue their fantasies of law enforcement and justify their behavior with after-the-fact claims of self-defense.

Justifiable homicides in Florida have tripled since the law went into effect. In 93 cases in which the defendant claimed self-defense under the new law, charges were dropped in 57 of them, and 7 others were acquitted by a jury. This law is irresponsible and should be repealed. If the shooter bias studies are to be taken seriously, the equal protection claise of the Fourteenth Amendment would seem to provide a basis for overturning them.

Kiss Kiss Bang Bang

Does anyone remember those Looney Tunes in which Bugs Bunny is fleeing from a speeding bullet, and every time he zig-zags, the bullet changes direction along with him? Well, Sandia National Laboratories, a division of Lockheed Martin, has announced the successful development of a rifle for use by the military that fires self-guided bullets. In theory, these weapons could hit a laser-designated target from very long distances. Just point and shoot. The bullet does all the work. Sandia's motto on their website is “securing a peaceful and free world through technology."

Photo by www.defenceimages.mod.uk

The bullet works on the same principle as a smart bomb. An optical sensor in the nose of the bullet detects a laser beam on the target and sends feedback to a control system, also built into the bullet. The bullet flies straighter than usual because it lacks the spin caused by the uneven surface which keeps a normal bullet from hitting its target. It has a smooth surface and fins to prevent it from spinning. According to Sandia, a typical bullet misses a target one-half mile away by 9.8 yards, while this bullet only misses by 8 inches.

The key to near perfect accuracy is “painting” the target with a laser beam. I presume the laser is built into the scope of the rifle. The shooter will only have to keep the laser on the target until the bullet arrives. However, this limits the range of the weapon. An alternative would be a two-person operation, with one person spotlighting the target with a laser while the other fires the weapon. In theory, this would allow the shooter to hit the target from an even greater distance and may not require a direct line of sight to the target.

Doesn't this begin to sound like a political assassin's wet dream?

Stanley Milgram and other social psychologists who have found that anything that decreases salience of a helpless victim increases willingness to harm that victim. The physical distance between an aggressor and his target is one factor determining salience. It's much easier to kill someone from a distance of half a mile than from three feet, where you have to see, hear and smell his death.

Here's a cheerful video from a fellow who doesn't see any downside to this weapon:

Sandia's press release says, “Potential customers for the bullet include the military, law enforcement and recreational shooters.” (That's funny. The reporter on the video said that too. I guess plagiarizing from a corporate press release is much easier than writing your own copy. And they don't seem to mind at all!) The only impediment would seem to be the cost. Of course, cost is no problem for the U. S. military, but every police force in the country will want to have them as well. Not to worry. I'm sure most communities will be willing to lay off a public school teacher or two in order to accomodate them. As these weapons become more common, how long will it be before they start to disappear from the supply room?

Certainly, many hunters will want to have one in order to add a few extra inches to their manhood. But it seems to me that self-guided bullets are the hunter's equivalent of “fishing” by throwing a stick of dynamite into a pond and picking up the dead fish as they float to the surface. They're just not very sporting.

What's to keep every moron in the country from owning one of these toys? The National Rifle Association will surely insist that anyone who can legally buy a gun can has a right to have one that fires self-guided bullets. After all, they are needed in order to protect ourselves from our neighbors, who we suspect are also planning to purchase one of these weapons.

Price may turn out not be a problem. Sandia says these weapons “can be developed inexpensively and rapidly.” You and I have already paid for the R&D. There are big bucks to be made by pricing them to sell. Public subsidies followed by private profits. This is the corporate welfare state as it's intended to function. 

If one of these rifles and six bullets winds up costing less than a laptop computer, I'm going to be worried.