Following a suggestion that I do a blog instead of using FB notes, I thought I would try it for a time. Here is the FB note that generated the suggestion, just to get things rolling:


Frankly, I'm just plain disgusted. I am disgusted by sensational journalism on the Sandy Hook murders but also I'm disgusted by people too lazy to question what they encounter in the news.

Given more than half a lifetime of doing hard science, I'm a very data-driven person. I question what I read and I try to verify facts on issues that matter to me. Given all the squeal in the news over the Sandy Hook shootings, I decided to hunt down some real statistics on gun-related deaths vs. other types of death. This is a brief description of what I found.

Data Sources and Limitations

If metadata on statistics bores you, then skip this section. After all, in most news reports or pundit pontification, sources of data and data quality are seldom discussed, probably because journalists try to keep their readers' attention - and most people seem to be turned off by math and statistics. But I'm not a journalist; I'm a science nerd who knows that data has to be validated before it is used.

There are some limitations on data available to my quick online search. After going through several different databases on the internet, I decided to concentrate on three sources: US CDC statistics on the leading causes of death in the USA, a study on uninsured deaths from the American Journal of Public Health (AJPH), and the US Justice Dept. homicide trends website.

The CDC reports on the leading causes of death for every year. The CDC website I used for leading causes of death did not report on data prior to 2008 and later than 2010. Looking at the data, I found that ranking the leading causes of death in the USA did not vary at all from year to year, though the absolute number in each category changed by statistically insignificant amounts. The variations between years was small enough that averaging them would not significantly improve those numbers statistically when compared to using any discrete year's data. The numbers I use here are from the 2009 report (http://www.cdc.gov/Injury/wisqars/pdf/10LCD-Age-Grp-US-2009-a.pdf). I picked 2009 because it dovetailed with the most recent CDC report on death and injury rates, which uses data also from 2009 (http://www.cdc.gov/nchs/data/nvsr/nvsr60/nvsr60_03.pdf).

Unlike the annual CDC leading causes of death numbers, the AJPH study averaged gov't death data from 1986 through 2000 (http://ajph.aphapublications.org/doi/pdf/10.2105/AJPH.2008.157685). So right off the bat, we're looking at CDC annual data vs. time-averaged datasets from the AJPH study. It's actually a minor difference in this case since the ranking of causes of death is very stable over time. There is one other way that the AJPH study differs from the CDC death stats. The CDC cause of death stats differentiate many different causes of death whereas the AJPH data differentiates only between insured and uninsured deaths for adults regardless of primary cause. Unlike the minor difference between discrete vs. averaged data, this difference is one we need to keep in mind and should approach with some caution when making comparisons.

The last set of data is from the Bureau of Justice Statistics at the Us Dept. of Justice. The data look at choices of murder weapon trends over several decades, from the 1970s through 2005. This is data that I would encourage everyone to look at for themselves because of large variations in firearm-related deaths that occur in the 1980s and 1990s (http://bjs.gov/content/homicide/weapons.cfm). The data trends over time are too complex to describe here in brief other than to say that most of the variations in the data are due to handgun use by juveniles and young adults. In comparison to handguns, trends in the choice of other murder weapon are relatively stable.

The Statistics

The CDC reports on the 10 leading causes of death every year, for the overall population and for different age groups. Using the 2009 numbers, the overall causes of death were, in order, heart disease (599413), Cancer (567628), chronic respiratory diseases (137353), strokes and blood clots (128842), unintentional injury (118021), Alheimer's (79003), diabetes (68705), flu and pneumonia (53692), nephritis (48935), and suicide (36909).

The age group break-outs are interesting. Most of the heart disease, cancer, respiratory disease, and cerebrovascular deaths are concentrated in the 65+ age group. In contrast, the story in school-aged deaths is quite different. For the age groups of 5 to 9 and 10 to 14, the first, second and fourth causes of death are unintentional injury, cancer and murder. The third leading cause of death for the 5 to 9 group was congenital anomalies, and for the 10 to 14 group it was suicide.

The fourth leading cause of death for both school-aged groups was homicide. That's a shocking and scary statistic if we look at just the cause of death rankings without looking at the actual death rates. Let's now look at those raw numbers and rates. The total number of homicides for 2009 was 16799 (http://www.cdc.gov/nchs/data/nvsr/nvsr60/nvsr60_03.pdf), which is 0.7% of all causes of death in 2009. There were 199 homicide deaths for the 5 to 9 group and 186 homicide deaths for the 10 to 14 group. Given an overall population of ~300 million in the USA, those are not big numbers when compared to the total population. One interesting statistic involves homicides vs. suicides. Compared to murders, there were more than twice as many suicides in 2009 (36909 suicides = 1.5% of all deaths), and this relationship applies to the 2008 and 2010 data also. There are twice as many suicides than murders in the USA.

Out of the 16799 homicides in 2009, 11493 (68%) involved firearms. Out of 36909 suicides, 18735 (51%) involved firearms. In addition to these, there were 333 firearm deaths due to law enforcement actions, 554 firearm deaths due to accidents and other unintentional acts, and 232 firearm deaths where intent could not be determined. Adding these up, there were 31347 deaths related to firearms in 2009.To wrap up our look at the CDC cause of death statistics, for the school-age group of ages 5 to 9, 53 (27%) homicides involved firearms. For the 10 to 14 group, 115 (62%) homicides involved firearms. These figures are lower on a per capita basis compared to the national percentage that 68% of all 2009 homicides involved firearms. In comparison, out of 259 suicides for the 10 to 14 age group, 64 (25%) involved firearms.

The numbers from the Bureau of Justice Statistics involve looking at choice of murder weapon, which is a more focused look at homicide than that available in the CDC cause of death statistics. Here are the numbers. First, more homicides in the USA involve guns than any other means of death. For example, in 2005, there were 11346 gun murders vs. 5346 non-gun murders. Restating that on a rate basis, 68% of all homicides in 2005 involved firearms. That number has not varied a gr eat deal since 1995: the annual murders with a gun percentage varies between 63% and 68% between 1995 and 2005, and the CDC 68% statistic matches well with these Dept. of Justice rates. To boil this down into a sound bite, based on gov't statistics, approximately two-thirds of all homicides in the USA today involve firearms.

It is a useful exercise to look at the Justice Dept. breakouts of their weapons stats. Overall, the numbers do not vary much between 1995 and 2005, so let's proceed with using the 2005 data for now, with the valid assumption that the 2005 data is representative of the previous decade's numbers as a rough approximation. In 2005, 50.8% of all reported homicides involved a handgun, 17.2% involved guns other than handguns, 12.9% involved knives, 4.0% involved blunt objects (e.g., baseball bats), and 15.1% involved all other weapon types. The real standout statistic here is the handgun data. More than half of all murders in the USA involve handguns. Looking at the handgun numbers overtime, it is accurate to say that more than half of all murders have involved a handgun since 1990. Before 1990, handgun murders varied between 43% and 50%. Handgun murders peaked in 1994 at ~58% and have slowly declined since.

Given the recent tragedy in Newtown, Connecticut, it would be useful to know what percentage of murders are committed with assault rifles, other automatic weapons and semi-automatic weapons. I make those distinctions because the news media did not. In the Newtown shooting, no assault rifles were used. The newsies got it wrong. The Bushmasters used in Newtown are classed as semi-automatics. The snarky side of my brain is muttering that if we want to go down the route of mandating firearm education in this country, we should require it of journalists as well as of gun owners. But my snarky side is not writing this little piece, so I will refrain from stating that opinion here. Such numbers probably exist out there already, and I suspect I might find them in the FBI's yearly compilations of homicide statistics, but I haven't found them so far.

There are some other interesting trends in the Dept. of Justice data, mostly on who commits the most murders and who are the victims of most murders. Homicides in the USA rose in the 1980s and then declined in the 1990s, and most of that variation involved gun-related homicides committed by male adolescents and male young adults using handguns. In comparison, murder rates for using other means by other age groups were relatively stable. In general, most victims of gun-related murders were young, involving older children, adolescents and young adults. Now the data are a bit more nuanced than I can convey here, so again, I urge people to go to the source of this information directly at http://bjs.gov/content/homicide/weapons.cfm, and read about these trends yourself. Don't let pundits feed you their already-spun data . Do your evaluation and exercise your own judgement.

The last data I want to present is the estimate of the number of deaths due to the lack of health insurance in the USA. The AJPH article is online for anyone to read at http://ajph.aphapublications.org/doi/pdf/10.2105/AJPH.2008.157685. This is very different data than the gov't death and homicide statistics. First, the data does not differentiate by other causes of death. A death from cancer or congestive heart failure for someone who couldn't afford a yearly physical counts the same as a death for a homeless person whose cold could not be treated until it became a pneumonia that could not be legally turned away by an emergency room. Most of these no-insurance deaths are concentrated in chronic illnesses that could have been treated or cured if the person involved could afford preventive care. This data is very different from the causes of death and homicide statistics. The data used covered patients aged 17 to 64 and tracked from 1986 through 2000. After a lot of sophisticated statistical analysis, the punchline of this study was that approximately 45000 deaths per year could be attributed to the lack of health insurance. In other words, every year some 45000 Americans die because they lack the financial means for non-emergency care like a visiting to a doctor's office or purchasing prescription drugs.

Playing Pundit

One of my hates in life is the new media circus. I personally found the reporting of the Newtown shooting as very much over the top. Interviewing traumatized and grief-stricken grade school students is not news in my book - it's sensationalism of the tackiest sort. And of course, since this was a shooting at a school, we have once again plunged into the caliopy of debate on gun control. The problem I see in this is a lack of perspective and the unfortuneate tendency to let the news media drive the discussion of issues nationally without any attempts at thoughtful analysis.

Despite the certainty that I will be portrayed as insensitive and disrespectful of the tragedy at the Sandy Hook School, I want to present a perspective on the subsequent debate on gun control that this shooting has rekindled.

Let's look at the numbers. Twenty-seven people died in Newtown, 20 chicldren and 7 adults. Using the CDC data, one can make a good case that approximately 2% of all homicides per year involved children aged 5 through 14. That's not a lot. Yes, the death of children is almost always more tragic and more heart-rending than the murders of adults, but compared to the number of adolescents and young adults who are murdered, it's nothing. So one could make the statement right now that we are in a debate over gun control over an age demographic that is not significant. Granted, it is quite disturbing that fourth leading cause of death for children is homicide, but the cause of death rankings are relative measures that do not take into account hard quantitative comparisons. What's a camparison? Here's one: the most common deaths in the USA occur in people over 65 from the common diseases involving the heart, lungs, and circulation. Basically, grandpa is more likely to die of a stroke or heart attack than any elementary school kid anywhere in the country. We're having a debate fueled by the news media that neglects to inform us that school kids hardly ever get murdered. Hence, we tend to treat these deaths not as the exception that they are, but as the rule.

Let's look at some of the other numbers here that we've not been reading about in the ongoing gun control debate. Right now, we're hearing a lot of talk about assault rifle bans. To be charitable, let's overlook the fact that reporters don't know the difference between assault rifles and semi-automatic rifles. When we go to look at weapons used in homicides, the murder weapon that really stands out is the handgun, not semi-automatic or assault rifles. More murders are committed with handguns compared not only to all other guns, but to all other murder weapons! And yet, where is the hue and cry over bans or better controls for handguns? Frankly, as a gun owner, I could live with stricter gun control on handguns because they are so screamingly prevalent as a murder weapon in this country. I'd be personally unhappy over restriction of semi-automatic rifles because my choice of bear and deer rifle is a semi-automatic; however, I could live with going back to my vintage Winchester 30-30 if semi-automatic weapons were banned. My point is that if we're going to seriously discuss removing a firearm type because of its use in homicide, we should really start with handguns, not mis-identified assault rifles. Gun control measures should be based in facts, not emotional knee-jerk reactions fueled by the sensationalist press. Handguns first, please.

There's been a lot of discussion about mental health care in relationship to the Newtown shooting. There's something about this particular debate that irks me, mostly because it's a no-brainer that mentally ill people shouldn't have easy access to guns. But there is another way to approach this issue. It's useful to ask if the murder rate for elementary school-aged children would be meaningfully lowered if mentally ill people were prevented from owning guns. The answer from statistics is "no" because the number of grade school children murdered with guns every year is so small that statistically-significant improvement in the child murder rate is mathematically improbably. Yes, that's an insensitive statement based in cold hard analysis, but when considering costly screening of all potential gun owners, we should heed the facts and not the impulses of knee-jerk emotions. Like it or not, gun control measures targeting mentally-ill adults would not have prevented the Sandy Hook School shooting.

I could keep going here, pulling numbers from different studies and looking for real trends in homicide that might be mitigated with intelligent gun legislation. It should be more than obvious that I dislike arguments based on knee-jerk emotional public responses fueled by sensational journalism. But I will wrap for now by looking at a different number altogether - the number of deaths attributable to the lack of health insurance, estimated at 45000 per year. That's more deaths than all homicides and more deaths than all suicides in this country. If death from lack of health insurance was a category in the CDC causes of death statistics, it would displace suicide as the tenth leading cause of death for the overall population of the country. That's a scary statistic.

Frankly, it is far more likely that you or I already know someone who has suffered the financial destruction of getting sick without health insurance than it is that you or I would know a child who was murdered. Approximately 1 in 6 Americans lacks health insurance right now. We all know people who can't afford it. I'm a small business owner and I can barely afford it. But less than a half of a percent of this country's population falls into a murdered child statistic. Yes, it is a heart-rending tragedy that 20 school kids died at the hands of a young man who used someone else's guns to kill them; but to drive a national debate on issues related to the death of a statistically insignificant portion of the population when more pressing matters exist that involve millions lacks perspective in my opinion. National policy should not be driven on gut reactions to protect our precious school children, especially when most of the suggested fixes would not have prevented the Sandy Hook shooting in the first place. National policy should be driven by finding ways to use our finite tax dollars to benefit society as a whole in the most financially-prudent way possible. Personally, I would suggest lobbying to make every member of Congress personally responsible for funding their own health insurance, or at least terminating their life-long health benefits after they step down from office.

What? You didn't know that election to the Senate or House granted most of our elected politicians health benefits for life? How about we strip them of their health benefits before we move on to matters of gun control. It certainly would save a few tax dollars...

Personally, I have the misfortune of knowing three murdered children. One was shot in his own home by a still unknown murderer. The other two were murdered by their own father who then killed himself by setting the house on fire. The murder of children is always tragic, no matter what the circumstances.

Please forgive my sins of any typos and misspelling. I may have a PhD in hard science but that doesn't mean that I can spell worth a bean.

Feel free to share this if you like what I had to say.