Lies, Damned Lies, and Statistics: Suicide or Homicide?

Warning: If you follow some of the links in this post, you’re liable to see some shocking things.

Two recent murder trials revolved around the issue of whether or not the victim’s “manner of death” was homicide or suicide: Phil Spector and Raynella Dossett-Leath. In neither trial were suicide statistics cited as evidence (thankfully), because statistics aren’t evidence. The only time statistics can make it into evidence, I believe, are when an expert witness uses statistics in forming an opinion about some aspect of a crime. (But this is also very tricky.)

Even so, outside the courtroom it seems as if everyone wants to talk about the statistics. In the Phil Spector trial, defenders of Spector claimed that gunshot wounds to the mouth were proof-positive of suicide. A comment from a reader of this blog claimed that “55% of gunshot deaths are suicide.” In the Dossett-Leath trial, the whole police investigation was predicated on the assumption that three-bullets could not possibly be fired by a suicide victim.

What do statistics say?

I decided to check out U.S. suicide statistics, to see if these beliefs were true.

Guess what, they are not. Not by a long shot (hideous pun intended).

Is it true that 55% of gunshot deaths are suicide?

  • Yes, according to the Center for Disease Control (CDC), in 2005, 55.4% of gunshot deaths were suicide; 40.2 % were homicide. However, statistically speaking, 51% of anything is random chance; and every statistical proof is subject to a “margin of error” (also known as the “standard deviation”)of something like plus or minus about 4 percentage points. So statistically in gunshot-caused deaths it is just barely possible to claim gunshot is an indicator of suicide (Clarkson); and it is definitely not possible to claim that gunshot is an indicator of homicide (Leath).

Is it true that a gunshot to the mouth is proof of suicide?

  • Not proof—only suggestive.
  • German forensic scientists conducted a statistical study to determine whether it is possible during an autopsy to distinguish homicide gunshots from suicide gunshots. The findings are significant. More than one gunshot wound per suicide occurred in 5.6% of the cases (including one case of 5 wounds). 7.5% of homicide wounds were near-contact (Spector). Suicide gunshot wounds occurred in the mouth in only 20% of the cases (Clarkson), and in the forehead in only 11% of the cases(Leath) (but, notably, they did occur in these locations). The conclusions were: “Consequently, some bullet path directions cannot be considered indicative of suicide: downwards and back-to-front in gunshots to the temple, left-to-right in gunshots to the left chest and downwards in mouth shots. The isolated autopsy findings can only be indicative of suicide or homicide but the combined analysis of several findings can be associated with a high probability.”
    • I believe the bullet’s path in the Spector-Clarkson case was downward into the base of the spinal column (hence indicative of homicide). I believe in the Leath case the bullet’s path was slightly upward, with flesh being embedded on the underside of a piece of the headboard (hence indicative of suicide).
  • A site called has some gory but suggestive information concerning suicide versus homicide wounds. Among the points of interest are that a suicidal man with a gun can, indeed, shoot several bullets before succeeding in killing himself; and homicide can, indeed, be committed by shooting someone in the mouth.

Is it true that suicide by gunshot is always fatal on the first try?

  • Absolutely not. Most suicide attempts are failures. Depending on the age group for which the statistics are calculated, the range of the ratio of attempts to suicide can be as high as 200:1 and as low as 14:1.
  • Emergency rooms see hundreds of thousands of failed suicide attempts each year. According to the CDC in 2005, 372,722 people were treated for self-inflicted injuries.
  • According to the National Institutes of Health, overall 8 to 25 attempts at suicide occur for every fatal suicide; most nonfatal suicide attempts are associate with drug use (as in Mr. Leath’s case)
  • All Experts posed an interesting question and posted an interesting answer about suicide and multiple gunshots. Here is part of what they said: “People who commit suicide with a gun typically will instinctively pull away as the trigger is pulled, and people have wounded themselves superficially by doing so.” This reminds me of the Dossett-Leath case in which the defense tried to prove that the first of three shots grazed the victim’s scalp. It also is suggestive with regard to Ms. Clarkson, who did not “instinctively pull away as the trigger [was] pulled,” in my opinion because the gun barrel was being rammed into her mouth by Spector;she didn’t see it coming and she couldn’t move her head.

Who is more likely to commit suicide? A 40-year-old woman or a 55-year-old man?

  • A 55-year-old man. Suicide attempts increase with age, after age 25. In other words, the highest rates of suicide are among very young people and very old people. (Not that I think 55 is very old, but Mr. Leath did have a history of dementia-like behavior.)
  • According to, white women in 2005 had a suicide rate of only 5 per 100,000, as opposed to 19.7 per 100,000 for white men.
  • According to, California is 42 out of 51 states (plus D.C.) in incidents of suicide (in other words as a Californian Ms. Clarkson was less likely to commit suicide than most other American women). Tennessee, however, is 18 out of 51, which means that David Leath was more likely than most American men to commit suicide. (BTW: Illinois is 43 out of 51. Despite the cold, we’re happier than the people in California.)
  • According to the CDC, in 2005 men were 6.8 TIMES more likely to commit suicide with a gun than women. Overall (statistics for more than one year), men are four times as likely to commit suicide than women, but women are two to three times more likely to attempt suicide. In part, this is because men use guns more often than women, and gunshot suicide attempts are more lethal.
  • According to the CDC, men are more likely to use a gun to commit suicide than women (almost 58% of male suicides are with guns). The most common form of suicide for females is poisoning (almost 40%).

Is it possible for a suicidal person to hide their depression from loved ones? In other words, do the families of suicides always suspect that suicide is a possibility?

  • Of course it is possible to hide suicidal feelings from loved ones. It happens all the time. The American Society for Suicide Prevention lists some warning signs to watch out for, though. An interesting one is “Giving away prized possessions,” which reminds me of Mr. Leath’s sudden desire to change his will before he died. However, acquiring things is not a sign of suicidal depression (as in Ms. Clarkson’s purchase of several pairs of shoes, which I can assure men always cheers a woman up. I don’t know what it is about shoes.)

Where do suicides occur more often? In the residence of the victim? Or in another location?

  • I gave up trying to Google this issue. It was too sad. A few weeks ago, I saw an I.D. Channel show on “Crime Scene Cleanup” in which a crime-scene worker told about a person who got up from the table at a family holiday meal and went into another room where he shot himself in the head.

From these statistics, you can conclude whatever you want about the Spector and Dossett-Leath cases. In fact, that’s why statistics aren’t often presented to juries.

Statistics are useful to scientific researchers since they can be used to prove or disprove hypotheses about general cases. But statistics cannot be used to prove that some specific “case” is true or false.

Let me repeat: Hypotheses about generalities are proved with statistics. Crimes cannot be solved with statistics.

Why? Because in every set of statistical data, there are always data points that prove the affirmative and data points that prove the negative and even data points called “outliers” that seem to be impossible. Take a look at the German study mentioned above. The medical examiners looked at a case of suicide in which the individual shot himself IN THE BACK OF THE HEAD.