Gun ownership prevents cancer?!?

[unixronin]

A couple of people I know have recently done OKCUpid's 'politics test'.  I haven't done it, because it offers no neutral/undecided/"it depends" answer choices, but I went looking to see if there was a link to compolain about that to the authors.  There wasn't, but while looking, I came across a "Death test" from Harvard that purports to predict your life expectancy with good accuracy.

At the bottom of the second page, at the bottom of a list of health risks such as a personal or family history of cancer, hypertension, obesity, etc, it asks whether you or your family have a history of gun ownership.

Excuse me?!?

"Gun ownership is not a medical condition, you assholes," I thought.  "You just threw your entire test into my 'Probably junk science and FUD' mental bin."  So I started doing a duplicate "control" test, and the third page, diet, has a checkbox for "Something you killed yourself," but no entries for grains, white or brown rice, non-exotic fish, etc.  OK, we're starting to get into too many variables here... I'm not controlling this FOUR ways.

(Side note:  The test has some really offensively flashy banner ads that AdBlock can't seem to block.  I'm about ready to just declare that all advertisers should be killed on sight.)

And "Do you often walk places you could drive instead?"  Come on, people, that's a STUPID question!  The other way round would make far more sense.  And then it goes on to a "reflex text" page that doesn't test reflexes at all, it tests hand-eye coordination and the responsiveness of your mouse and web browser.  It asks you to click eight scattered checkboxes with the mouse.  I can't consistently repeat it multiple times within 2 seconds of each other, and I totally roll to disbelieve that any normal human can repeatably do that page in their posted best time of 1.4 seconds.^[1]  Hell, the checkboxes don't even consistently check on the first click.

The end result:

Interestingly, checking the gun ownership choices seemed to have no effect whatsoever on the outcome of the test.  Both my gun-ownership test, and my no-gun-choices control, predicted I will die in November 2044 at an age of 84.3, and both offered the same first four predicted causes and probabilities of death (in ascending order of probability):

• Wounds: 2%
• "Drowning of the lungs": 5%
• Loneliness: 13%
• Car accident: 24%

Oddly, the test in which I checked the gun-ownership options gave my first predicted cause of death as "Heart attack: 56%", while the test in which I AVOIDED them gave me a 56% chance of cancer instead.

So this seems to be saying that according to Harvard, gun ownership prevents cancer ..............

I can tell you for free, given the choice between dying of a heart attack and cancer, I'll take the heart attack.  :)

---

[1]  Actually, I'm guessing whoever got 1.4 seconds was able to do it by just hitting tab, select, tab, select, tab, select.....

---

(Crossposted to guns)

Comments

[therealocelot.livejournal.com]

"You just threw your entire test into my 'Probably junk science and FUD' mental bin."

But..but... I thought OKCupid was the height of accuracy and appropriate use of scientific methedology!

On a more serious note, it's possible that various factors interact - gun ownership influence cause of death, but gun ownership and suicidal tendencies does.

[unixronin.livejournal.com]

That's a definite possibility. Folks who attempt suicide with a firearm are far more likely to succeed.

[beckyzoole.livejournal.com]

I recall reading from a respected source -- but now I cannot remember exactly what that source was -- that states that introduced strict gun control laws found no real change in either the crime rate or the death rate... but that the suicide rate did go down.

A mandatory three-day delay for gun purchase was particularly effective in reducing the suicide rate.

[unixronin.livejournal.com]

This is a claim that's been made more than once. The truth of it seems to be that suicides using a firearm went down slightly, but people sufficiently determined to use a gun usually just found alternative methods.

[beckyzoole.livejournal.com]

I don't doubt that the number of suicide attempts would stay the same. But a firearm is a much more effective way of killing oneself than taking a handful of pills or other popular methods. A delay gives people time to think it over and back off the immediate urge to buy a gun and shoot themselves in the head because they got dumped or fired. Sure, very determined suicidal people will always find a way. But many impulse-suicides are thwarted.

The interesting thing is that most studies find no reduction or increase in the crime rate when handguns are controlled. Those who oppose gun control would expect the crime rate to go up when law-abiding citizens can't protect themselves as easily; those who support gun control would expect it to go down when criminals can't get their hands on guns as easily. But the crime rate continues to be affected more by the size and efficiency of the local police department than by anything else.

[unixronin.livejournal.com]

On the latter, I'll just reference the Lott study (the most complete and comprehensive conducted to date) and leave it at that. I haven't ever felt morbid enough to study the former in deep detail.

[beckyzoole.livejournal.com]

Do you have a URL for the Lott study? I've read references to it, but never looked at the whole thing.

[unixronin.livejournal.com]

I don't have a URL to the study, and don't know if the full study is available on the Web. I have a copy of the first edition of the book he wrote documenting it and based on it. His study was based on a county-by-county study of the entire US over a thirty-year period, excluding only a few counties for which he could not obtain data, and fully controlled for every other factor he could think of that might affect it. It was subsequently criticized by two San Francisco lawyers, using some very questionable statistical methods to challenge some of his findings; in the course of refuting their challenge, he actually did find one error in his work, which, when corrected, actually showed the data more strongly supported his conclusion than before the correction. No-one else has contested it. He's subsequently expanded and updated his book, and - I think - extended his study period to include newer data as it has become available.

If memory serves, he actually set out in the first place intending to prove that gun control worked, but the data from the study was so overwhelming that he was forced to change his view. I might be thinking of a different study by a different researcher there, though.

The capsule summary of his results was that across the US, where it is easier for law-abiding citizens to carry personal arms, crime is lower, and where restrictive laws make it harder, crime is higher. He further showed that the chronological relationship between changes in laws and changes in crime is such as to show that it is the change in law that is the cause, and the change in crime rates the effect.

(The most clearly I've ever seen this myself, incidentally, is in California's waiting period laws. California's waiting period for handgun purchases was first three days, then was increased to five, then seven, then ten, then fifteen, and in the two to three years following each extension of the waiting period, violent crime took a big jump before stabilizing at a new level, then jumped again right after the next extension.)

[unixronin.livejournal.com]

Oh, and to be fair, the test does supposedly come out of Harvard, but oddly, it's not from the school of medicine, it's from the math department. One presumes its basis is therefore actuarial rather than medical.