🎓 Dr. Freemo 🇳🇱 @freemo@social.cleverlibre.org

@sturgman I would be happy to next time im by a computer and can dig it up. The graphs are from an old post so not sure where the source is off hand.

@rrb

In social sciences there is no perfect analysis. But this is still superior to simple corelation in this case. That said there are more advanced techniques that are arguably even better. But those are hard to explain to people.

@protecttruth

@sturgman

The truth is, this is my field, I am an expert in it.. we can disagree on the conclusions of course, Im not saying you have to share my conclusion by any means.. but I can say it absolutely is in line with how granger tests work.

When i say you need enough data points it has nothing to do with the length of the timeline. The number of data points you care about is how many times the gun laws have changed. So for a single chart its only counting as a single data point.

And to your credit your questions are very valid. Nothing is proven from a single chart in isolation, of course. The problem is when you do this same analysis over many dozens of examples the results all turn out the same, which becomes very telling.

@AndyLowry @protecttruth

@JonKramer

Unfortunately that sort of analysis is almost useless simply because you have no way of controlling a rather long list of unknown confounding factors. Its one reason why scientists avoid this very sort of analysis... well, if they are serious and objective anyway.

@protecttruth

@sturgman

Then it seems you misunderstood the article.

To get a proper result from the granger test youd need to run it across a larger dataset of this nature for sure. But the images here would be individual examples in that regard of it passing the granger test.

@AndyLowry @protecttruth

@sturgman

No thats not how granger causality works. It doesnt need to be shown as seperate time series. But if you orefer to think kf it that way you can think of the tine series representing changes in gun laws being the other time series (which in this example is just overlayed onto one)

@AndyLowry @protecttruth

@sturgman

Of course its incomplete, this is social media afterall. It took me a life tine of study to become an expert on this subject, so that cant be conveyed in full over social media. But yes it does test for predictive causality. Which is a fancy way if sayibg does the existance of one event (gun law changes) predict another event (change in homicide). This is usually what we mean when we talk about causality.

@AndyLowry @protecttruth

@rrb

As long as they are comparing a state against itself both pre-law change and post, and they do that analysis correctly, then yes that would be fine science.

@protecttruth

@rrb

I think if your beating your wife or husband then you should go to jail. Sinc eit is a violent crime and if you served jail time then I would be fine with them having their gun rights revoked. More generally anyone who has went to jail for a violent crime probably should be denied a gun.

@protecttruth

@rrb

When forcing suicidal people to wait because you want to dictate what they can or cant do with their body and when.. at the expense of a woman not being able to protect herself against a rapist... Id lean towards letting the woman protect herself.

Let it be an opt-in thing.. everyone at the age of 18 must register witht he government if they want to be forced to wait for a gun or if they want the privilage not to.. Then if they want to be protected against possibly being suicidal they now have that protection. For those who dont want the "protection" of a wait then they wont have it.

@protecttruth

@AndyLowry

Its probably less about embarassment and more about wanting to appeal to the masses.

@protecttruth