categories

archives

meta


"walls of the city" logo conceptualized by Oleg Volk and executed by Linoge. Logo is © "walls of the city".

graphics matter, year the fourth, updated, again

Thank God for my helpful readers.

Last time around, thanks to commenter TS, we were able to integrate importation numbers into the "graphics matter" series of posts, which only served to perforate the "more guns = more deaths" hypothesis even worse. Unfortunately, though, Shooting Industry Magazine rearranged their site sufficiently that some data was apparently lost, and I had to drop the 1981-1985 section of the graph due to not being able to adequately source my numbers; I will leave the wholesale fabrication of statistics to the "gun control" extremists.

shootingindustrymagazinefirearmproductionThis time around, Hoplophobic Healer reminded me that the WayBack Machine does exist, and after digging around its guts for a few minutes, I was able to find this archived page documenting firearm production from 1982 until 2001 (said page screencaptured to the right, just in case it tries to disappear too). Unfortunately, the Shooting Industry Magazine did not track imports that far back, but at least we have domestic production numbers again, and based on that, I have decided to go ahead and add the 1981-1985 section back to the graph.

For reference, the 1981 year was chosen for the lower bound of this graph simply because WISQARS Injury Mortality Reports do not go any farther back. I know there are sources that do go farther back, but then you run into the question of whether or not they track their information the same way WISQARS does, and that just throws this whole graphic into question. (And, yes, I am well aware that WISQARS changed how they code their data between 1998 and 1999, but if you look into the explanations of what all they changed and why, none of it really affects the numbers we are looking at.)

So enough of this jibber-jabber, right? As with before, all previous disclaimers, details, and other important stuff still apply, including using the Shooting Industry Magazine as a source again. Now, on with the show:

americanpopulationfirearmsdeaths2009-3

I am again going to dispense with the acres of text, and instead consolidate it all down to this disclaimer: you CAN compare the correlation coefficients in this dataset to the correlation coefficients in the previous 1981-2009 datasets, but NOT the 1986-2009 one I put up last week. Y’know, just to make things more confusing. Things, of course, did change with the re-inclusion of those five years, which is why I always maintain that more data is more better:

The raw number of firearms in America correlated to the raw number of firearm-related fatalities with a coefficient of -0.41741, which is a little stronger than before including the firearm importation numbers.

The rate of firearm ownership in America correlated to the rate of firearm-related fatalities with a coefficient of -0.80373, which is significantly stronger than before.

In other words, what I said before only continues to hold true: the hypothesis "more guns = more deaths" cannot be true in the fame of reference of American society over the past almost-three decades.

Again, unlike the anti-rights cultists who are currently weeping into their bourbon at the sight of those numbers, I am more than happy to show my work – after all, if I did not, I would never have you kind people to thank for filling in the holes I made.

16 comments to graphics matter, year the fourth, updated, again

  • I can only wonder when they will start claiming to have debunked you.

  • [...] Or rather I’ll point you to the excellent work Linoge at Walls of the City blog has done in debunking the idea that ‘more guns = more crime’. Convince me we need to stop putting more guns on street if you can. I’ll read what ever [...]

  • [...] because honestly it is the only thing our opponents in the gun control debate can argue with.  Facts and statistics are not on their side and we know it by the rants about “blood in the streets” that never actually seem to [...]

  • Well, the Brady Bunch has already attempted to misrepresent this graphic as “proof” that the Brady Act “worked”… does that count?

  • Any interest in adding a line for FBI UCR national violent crimes rate? That’s at 40-year lows, after all, and the “more guns = more crime” meme is still out there.

    http://www.ucrdatatool.gov/Search/Crime/State/StateCrime.cfm

    Sadly the online UCR query tool doesn’t allow slicing by “weapon used” to produce a nice year-by-year table; the violent crimes weapon used data is available, except for forcible rapes, but it would be a bit of work to collate it all from the individual reports. For instance:
    http://www.fbi.gov/about-us/cjis/ucr/crime-in-the-u.s/2010/crime-in-the-u.s.-2010/tables/10tbl19.xls
    http://www.fbi.gov/about-us/cjis/ucr/crime-in-the-u.s/2010/crime-in-the-u.s.-2010/tables/10tbl20.xls

    That would allow adding another line, FBI UCR national violent crimes using a firearm rate, so the “more guns = more gun crime” meme could be addressed too.

    Or would that be getting too cluttered?

  • *cough*

    The FBI has not released the finalized numbers for 2011 yet… should be sometime in September.

  • D’oh! Me shut up now.

  • …well, not quite. The individual UCR reports online only go back to 1995, which limits the scope of the “more guns = more gun crime” meme graph, but the overall UCR query tool goes back to the ’60s, so you could add a non-firearms-specific violent crimes rate line for the full graph back to 1981, which would address the “more guns = more crime” meme better.

    Heh. I keep harping on this. Sorry.

  • Motor-T

    This is a great resource. Would it be possible to break firearm related deaths down to, suicide, accident, unlawful homicide, and justified homicide? Or is the reporting on those too varied state to state?

  • @ John Hardin: Yeah… honestly, that might require too much effort out of me ;). If you ever get bored, though… ;)

    And, strangely, that 2010 graph you point to doesn’t break out weapon-assisted rapes, which kind of throws the data right off the bat. Odd.

    @ Motor-T: Nah, the CDC breaks out all of those as individual things, and if you want, you can see them separately. This graph was to address the overall, fallacious notion that “more guns = more deaths”… not to mention trying to break out all of those on the same graph would make it really incomprehensible ;).

  • Linoge wrote:

    @ John Hardin: Yeah… honestly, that might require too much effort out of me. If you ever get bored, though…
    And, strangely, that 2010 graph you point to doesn’t break out weapon-assisted rapes, which kind of throws the data right off the bat. Odd.

    Yeah, I saw that about the rapes too. I don’t understand the omission unless that datum isn’t collected in the first place.

    The underlying raw data sets are available so a “violent gun crimes” dataset might be possible would but a bunch of work to compose. I took a quick look at the list of data sets but haven’t looked at the raw data itself yet.

    However, the overall violent crime rate data (ignoring weapon used) is available formatted into a easy-to-import multi-year table via the UCR data query tool at http://www.ucrdatatool.gov/Search/Crime/State/RunCrimeTrendsInOneVar.cfm – adding the overall violent crime totals or rate is a matter of cut-and-paste.

    I played with the chart a bit, but the Y-axis ranges don’t work well – the rate per 100k (which is what’s usually quoted) is off the bottom of the graph, the rate per 1M crosses the bottom of the graph and rate per 10M is off the top of the graph. It looks nice – and is more impressive than the firearm-related deaths rate – if you do rate per 5M, but that’s a really odd scaling factor…

  • Well, remember that we can always screw with the multipliers for the other data as well… And, yeah, I would very much prefer to keep it to whole orders of magnitude, just for simplicity’s sake.

    I will look into it sometime this week, though I make no guarantees :).

  • Tom O'B

    I was playing with Excel and tried changing number of shall issue states to the total population in shall issue states.
    No surprises but reenforces increase in population doesn’t lead to shoot outs on main street.
    My version can be seen at home.wavecable.com/~obanntl/populationfirearmsdeaths2009withcrime-3-3.xlsx

  • I will have to look into that; thanks for doing the crunching. The only catch, of course, is that a state’s population is not necessarily indicative of how many people in that state have carry permits. Neither is a number of states in the union indicative of how many Americans in general have permits, I suppose, so it could work either way.

  • [...] as being rational and the only way to curb gun violence.  The thing is, statistically gun control has been proven to be ineffective over and over again.  Further when you compare nations with strong gun controls it becomes obvious it creates a [...]

  • [...] let us talk about that correlation. Oh, wait, I already did: The raw number of firearms in America correlated to the raw number of firearm-related fatalities [...]