Monday, February 15, 2016

Bad Science

http://www.realclearscience.com/journal_club/2016/02/02/study_of_171_countries_finds_link_between_public_mass_shootings_and_gun_ownership_109526.html
The United States now has more guns than people, according to estimates published last year in the Washington Post. But even before that disconcerting nugget of information entered the public realm, the U.S. still housed a rather obscene amount of firearms

Disconcerting?  Obscene?

Yeah...  it seems like you want to ban guns.


"Here's what we do know. After peaking in the 1980s and early 1990s, crime has plummeted in the United States. The rates of forcible rape, murder, violent crime, property crime, and aggravated assault are currently as low as they were in the 1960s."

So.... more guns don't cause more crime?  Any expanding on how that undercuts the core reasoning behind gun control?   None?

Ah, gonna focus on the mass shootings?  Right.

Lankford then explored how the number of mass shootings per country were associated with each country's homicide rate, suicide rate (used a rough proxy for mental health), and firearm ownership rate. While he found no link between the number of shootings and suicide or homicide rates, he found a highly significant (p<.01) link between the number of shootings and firearm ownership rates. 
Apparently, yes,  there's no link on homicide and gun rate but there is on the subset of homicides.

Oh-kay.  No pondering why that's happening?


Lankford noted a number of limitations to his study. Older incidents occurring further in the past and in countries without streamlined reporting systems may have been missed. Moreover, since public mass shootings are rare, the sample size is small for the forty-six-year study period.

So, a small sample size,  combined with bad reporting.    You know,  when you have a correlation that seems anomalous and is in part due to a small sample size,  there might be something off with that correlation.


But why let that get in the way when there's obsecene and disconcerting guns to ban!


Lankford also made clear that he utilized the definition of public mass shooting from the NYPD's report. The attacks "must have (a) involved a firearm, (b) appeared to have struck random strangers or bystanders and not only specific targets, and (c) not occurred solely in domestic settings or have been primarily gang-related, drive-by shootings, hostage-taking incidents, or robberies."

Which means despite the breathless writing earlier in the article San Bernadino wouldn't count.


He did however, state the natural conclusion from his findings.
"Perhaps the most obvious step the United States could take to reduce public mass shootings may also be the most politically challenging: reduce firearms availability."
Lankford noted that the approach seemed to work in Australia. After a public mass shooting in 1996 that left thirty-five people dead, the country's government passed comprehensive gun control legislation. 

Wait... so overall gun deaths's aren't correlated... but the natural conclusion is gun control?
Color me surprised.

But nobody wants to ban guns.  Do note they don't even say that Australia had mass bans and confiscations.

I also wonder which other rights and possessions these folks think should be banned because a correlation was found.  Is that how science works now?

"Hey we found a correlation between autism and visits to the doctor!  Let's   ban doctors!"

No comments: