SoylentNews is people
SoylentNews
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.
(Score: 2, Interesting) by Anonymous Coward on Thursday May 23 2019, @06:11PM
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.