Stories
Slash Boxes
Comments

SoylentNews is people

posted by janrinok on Thursday May 23 2019, @02:39PM   Printer-friendly
from the urban-rural-divide dept.

Swiss voters on Sunday approved a measure to tighten the Alpine nation's gun laws, bringing the country in line with many of its European partners despite the objections of local gun owners, Swiss media reported, citing official results.

Switzerland's public broadcaster said more than 63% of voters nationwide agreed to align with European Union firearms rules adopted two years ago after deadly attacks in France, Belgium, Germany and Britain.

The vote Sunday was part of Switzerland's regular referendums that give citizens a direct say in policymaking. It had stoked passions in a country with long, proud traditions of gun ownership and sport and target shooting. Switzerland, unlike many other European nations, allows veterans of its obligatory military service for men to take home their service weapons after tours of duty.

The Swiss proposal, among other things, requires regular training on the use of firearms, special waivers to own some semi-automatic weapons and serial number tracking system for key parts of some guns. Gun owners would have to register any weapons not already registered within three years, and keep a registry of their gun collections.

https://www.usatoday.com/story/news/world/2019/05/19/tighter-gun-laws-appear-pass-switzerland-despite-opposition/3731629002/


Original Submission

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: -1, Flamebait) by Anonymous Coward on Thursday May 23 2019, @03:50PM (7 children)

    by Anonymous Coward on Thursday May 23 2019, @03:50PM (#846676)

    Astronomy has no controlled experiments but doesn't have the same issues. I'd say its more that people studying public health "don't like math", so they are incapable of making any sort of quantitative predictions to test their theories.

    Starting Score:    0  points
    Moderation   -1  
       Flamebait=1, Total=1
    Extra 'Flamebait' Modifier   0  

    Total Score:   -1  
  • (Score: 5, Interesting) by ikanreed on Thursday May 23 2019, @03:58PM (6 children)

    by ikanreed (3164) Subscriber Badge on Thursday May 23 2019, @03:58PM (#846678) Journal

    When you typed this up, did you ever ask yourself "Am I making a meaningful point?"

    Of course people studying public health like math. That's what public health is: applying rigorous mathematical analysis to health trends. And most of these papers fit a fairly clear hypothesis, observation, conclusion paradigm.

    • (Score: 0, Redundant) by Anonymous Coward on Thursday May 23 2019, @04:05PM (5 children)

      by Anonymous Coward on Thursday May 23 2019, @04:05PM (#846683)

      Plugging stuff into SPSS/SAS isn't rigorous mathematical analysis. I mean real math where you start with assumptions and then derive a quantitative model that you believe may be more or less true (there will of course always be "perturbations").

      Something like 99% of that field is just misinterpreting the coefficients of arbitrary linear models. Change the model, change the coefficients. If you include interactions you can even get stuff like a positive "effect" of guns when male/female is coded as 1/0, but negative "effect" when you use 0/1.

      • (Score: 5, Informative) by ikanreed on Thursday May 23 2019, @04:21PM (3 children)

        by ikanreed (3164) Subscriber Badge on Thursday May 23 2019, @04:21PM (#846693) Journal

        Yeah, I can totally see wanting a non-linear model on this dataset [aphapublications.org]

        Stop shitting out ad-hoc justifications for why the evidence and analysis is wrong, and if you're not completely and utterly full of shit posit a counter hypothesis and model that better explains the data. Or fuck off. One of those two. Not this lazy pseudointellectual shit.

        • (Score: 2, Interesting) by Anonymous Coward on Thursday May 23 2019, @06:11PM

          by Anonymous Coward on Thursday May 23 2019, @06:11PM (#846731)

          Sorry but you are not following at all.

          1) The y-axis says "adjusted" suicide rate. These adjustments are probably including other variables in a linear model.
          2) There could be some third variable that correlates with both suicide rate and gun ownership, if you further adjust for that this plot would look totally different. The variables included in such models are inevitably a matter of convenience, which is why these coefficients and plots don't mean anything. It all depends on what you include in the model.
          3) I would definitely use a non-linear model for the purposes of prediction on such data. There is still quite a bit of scatter there, and you can probably do better than the linear model (but sometimes it isnt worth the effort).

          As I said, 99% of public health research amounts to making this exact same error of interpreting coefficients of arbitrary statistical models meant only to be used to predict some outcome. And the issue is hardly limited to public health research. Scary to say, but random anon person on the internet knows better than all these PhDs, it is true.

        • (Score: 0, Insightful) by Anonymous Coward on Thursday May 23 2019, @06:20PM

          by Anonymous Coward on Thursday May 23 2019, @06:20PM (#846734)

          Also, "linear" does not mean fitting a line. That may be another place you are confused.

          https://en.wikipedia.org/wiki/Linear_combination [wikipedia.org]
          https://en.wikipedia.org/wiki/Nonlinear_regression [wikipedia.org]

        • (Score: 1, Insightful) by Anonymous Coward on Thursday May 23 2019, @07:24PM

          by Anonymous Coward on Thursday May 23 2019, @07:24PM (#846750)

          This guy has a really good analysis on all these graphs:

          http://freakoutery.com/2018/07/everybodys-lying-about-the-link-between-gun-ownership-and-homicide/ [freakoutery.com]

          and especially the tricks related to getting a line through the origin as in your pic:

          http://freakoutery.com/2018/12/lying-with-gun-data-again/ [freakoutery.com]

      • (Score: 1, Troll) by Anonymous Coward on Thursday May 23 2019, @04:21PM

        by Anonymous Coward on Thursday May 23 2019, @04:21PM (#846694)

        Same AC. Try this:

          set.seed(12345)
          treatment = c(rep(1, 4), rep(0, 4))
          gender1   = rep(c(1, 0), 4)
          gender2   = rep(c(0, 1), 4)
          result    = rnorm(8)

          summary(lm(result ~ treatment*gender1))
          summary(lm(result ~ treatment*gender2))

        "Treatment effect" is 1.17 vs -0.38. That is just a particularly egregious example though. In general unless you believe your model is "true" the coefficients have no meaning. You can still use an arbitrary model to make predictions, but usually you would just use gradient boosting or neural networks for that these days.

        Now think about all the damage these people have probably done to society from drawing conclusions about pretty much random numbers.