Rather, when you sit in dismissive judgment on an approach without actually ever getting around to invalidating it, you are neither establishing your position as a fit judge nor actually managing what you wave an off hand at.
Agreed, if such a one does not subsequently follow their hand, which I did do.
In the sense that you believe it to be practically without value, I suppose. Otherwise, no.
In the sense that it was factual. If you mean by this my estimation of the value of the study is that it is practically without which, then yes I did make that case rather persuasively. It was a practically useless study, showing the biggest factor in explaining the greater number of people proportionally murdered in mass murders instead of just as individuals, is what year it was. That shows a statistically significant upward trend, which I never denied wasn't very important, but the study itself did nothing to explain the trend, beyond confirming that it is there.
It's a very important thing that it's there, and for that I suppose the study can be at least commended for that. But that was ancillary to what the authors of the study are trying to argue, which is that we should 'bring back' an "AWB," which their study does nothing to support.
So in as much as they published results, they are to be commended, but only by the accident of publishing something important, which they themselves overlooked in trying to argue unsuccessfully due to poor statistical reasoning their political view, which is entirely unsupported by any of their doctoral degrees, even if their statistics were sound.
There you swerve from the objective into the subjective valuation, but I'll come back to this in a bit.
And having already implicitly acknowledged this, I supported by claim with valid reasons. You act like I'm just stating opinions.
That would be the independent variable.
Right. The one independent variable was whether or not the "AWB" was in force, or not. Binary. 1 or 0. On or off.
It's always perfectly fine, mind you, to analyze data like this with just one variable, that's totally fine, perfectly OK. But when the explained variance of that one factor is such a low number that you don't even both to in the abstract explicate what it is, and instead offer that the year explains over a third of the variance (which is fine, for one factor to explain this much variance, but as I said, we're really hoping for models that explain something like 90% of the variance in the response; this is aside from the fact that 'year' is just not an actionable variable---it's 'independent' only because it's on the right side of the analysis, while the response is on the left), and leave out how much variance "AWB" specifically explains, we know through inference that "AWB" explained very little of it; much less than a third of it, iow.
The dependent variable being fatalities by gun.
I thought that it was all murders, not just homicides by gun, and I also thought that it was not just 'fatalities' but intentional homicides that they were looking at, but I could have missed it.
I don't want to say that I 'don't care about suicide,' because I do care very much about it as a personal tragedy, but I do not care about suicide as it might concern an idea regarding gun control, as I do not think that suicides ought to inform laws regarding rights. If anything, we have a right to not have one of our kin commit suicide, and I support things that discourage it, but not gun control.
For those following along, the intercept can be thought of as a constant and is the expected mean value of Y when all X = 0. X is the predictor. If X never equals 0 then it's meaningless, since its only the expected mean value of Y.
Then by your own admission, it's not meaningless at all in this particular case, which is why I mentioned it. Whether the "AWB" is or is not in force, the 'expected mean value' of annual murders in the US is in the neighborhood of 15000.
Miniscule is a subjective term.
'Miniscule' isn't a term at all before this thread. If you mean "minuscule," then again, I addressed that it was 'subjective,' and I supported my contention.
It's statistically significant
So another word here about statistical significance.
Before we even talk about 'coefficients,' let's just talk about statistical significance. We don't need to mention coefficients at all, and just deal with the p-value.
The p-value, when comparing two sets of data with each other, is the chance /probability that they are random wrt each other.
It's like flipping a coin forty times, and comparing the heads and tails of the first 20 flips, with the final 20 flips. The result will be that the p-value is higher than 0.05, which means that conventional statistics rejects the theory that they are correlated. There is a chance that the p-value will be 0.05, but it's a 0.05 chance that it will be 0.05. So in this unlikely case, the analysis will give us a false positive, saying that these two data sets are correlated, even though we know that they are not.
This goes to the understanding that statistics themselves are never the last word on any matter, we also need experts to weigh in on the logic that explains the statistics, when the statistics suggest a correlation. Statisticians provide raw material to subject matter experts.
When statistics do not deny the null hypothesis, then statistics provide useful and actionable information without any experts weighing in. It is when statistics show a correlation that then this information is submitted to experts for interpretation.
In this case, where the independent variable /factor is "AWB," we need experts to interpret exactly what this means, because it being a law, we are putting a lot of faith in that it being in force, or not being in force, influences the behavior of mass murderers.
and the outcome is a contributing argument for making the laws.
It is certainly not that. Before anything is done a much better study must be undertaken, unless this study is that study, and they analyzed factors that statistics ruled out, and if we know that, then that is valuable information, to know which factors positively do not affect their chosen response variable. If they did not cast a very wide net before performing their regression, then that must be done before I would agree with any change, especially in the direction of more gun control, instead of less.
I've already made the case that we two disagree on what The Right to Bear Arms actually means, and while anything like another "AWB" does not conflict with your meaning of the Right, it does conflict with mine, and for me there must be a much stronger case to make more gun control laws before I could consider the effort anything but wrongheaded, mainly because it directly conflicts with what I mean by The Right to Bear Arms, to say nothing of this particular study's conclusion being very weakly supported.
Though remember, I'm calling for models resembling our European cousins and those are much more encompassing
I know. The federal "AWB" that the study in question concerns, was about features of AR and AK patterns (etc.) and magazine capacity, and also did not involve any confiscation /compulsory buybacks. It was a spectacularly ignorant law.
, and the data much more compelling than this problematic side bar.
It is not. Statistically what you've done is select a subpopulation according to a nebulous set of conditions, that could instead be included in a more expansive analysis through the inclusion of a variety of binary independent variables. Based on what I recall you saying, you would include binary factors such as 'democracy,' 'European,' 'Western,' etc., each of which being either a 1 or a 0, and then they would all be combined into a single model, that wouldn't exclude any countries for which we have the same data. This would avoid your 'cherry picking' error entirely and produce a far more robust statistical model to explain your chosen response variable.
I focus on the one response variable that to me means the most, and that's murder. To me, if there are 10 times more murders in country B than in country A, even if all country A's murders are mass murders, country A is still far safer than country B. But from what I'm reading you say, you'd rather have more individual murders, just so long as we avoid mass murders, and I disagree with your values there.
Percent is percent.
I deny that's insignificant. And I think you would too, if pressed on the point of what we're actually talking about in terms of outcome and how that translates over larger populations.
If we weren't running headlong into what to me is blatant and flagrant infringement of The Right to Bear Arms (my meaning, not yours), then I could at least sympathize, but I cannot do so given the circumstances and all of them.
Okay. I can accept it, but it was oddly placed and seemed more a part of an overkill in establishing credibility for an easy dismissal that I still reject as an established conclusion, supra.
'Supra' seems more a part of an overkill in establishing credibility for an easy dismissal that I still reject as an established conclusion.
Now that was a neat rhetorical trick, the inference in copy/paste. I did supplement my approach, but if I'm talking about an abstract, the best way to present it clearly is to set it out, not spend time rewriting it and risk missing or mistatting something of importance. As someone familiar with the application of statistics to a number of models, but not someone who does this regularly or has a particular distinction within that field of study, it seemed the prudent course of action.
It was fine for you to copy-and-paste the abstract of the study, and it was fine for me to poke holes in it too.
We may need a harder and longer look at the full study.
Because I could analyze their same data myself, I personally would just want to see that data. As I said, if there are factors that they did include in their study, but that were ruled out, then that too would be valuable information for everybody to see.
There's something else worth mentioning here, and that is the notion of 'degrees of freedom' in a statistical analysis, and mainly I think it could boil down to this: You 'buy' degrees of freedom with more response data, and you 'spend' them by including more independent variables in the analysis. This study appears to use 37 data points in the response, which limits the number of independent factors that can be analyzed to something less than that number. I cannot recall how many are needed as a bare minimum, but it's safe to say the limit is something like 34. That's not just the independent variables themselves, but every possible transformation done to them also, meaning that if you include a square term that's another degree of freedom 'spent,' and the same for every interaction term; each additional one 'spends' another degree of freedom.
If you have just for example three independent variables (A, B, C), that's three degrees of freedom. If you include every interaction term also (AB, BC, AC, ABC), that's four more degrees of freedom, so just the three factors alone in such an analysis would 'spend' seven degrees of freedom, in a factorial model. Adding a fourth factor D would increase one more degree of freedom for D's main effects, plus another seven for the additional interaction terms (AD, BD, CD, ABD, BCD, ACD, ABCD), increasing the degrees of freedom required to 15 with just four independent factors in a factorial analysis.
A fifth factor would throw out a full factorial as possible, because it would require in total more than the 34 degrees of freedom that the 37 response data points offer.
So all that's to say that it's better to have more response data to perform an analysis that captures all possible factors. I would go further back in time than this study's analysis, instead of artificially trying to show an equal number of years before and since the "AWB," because the results would be more meaningful. The regression handles binary factors just fine, no matter how far back in time we go. Along the way, we could include other factors to help explain if possible the murder rate better than this study does.
You don't really need more data if you have enough to establish the significance from data mined pre and post and the variance during.
I would argue that you do, if you're not trying to hide something. What's the harm in starting back in 1917, for instance, or at least sometime before 1981? If the "AWB" is actually explanatory, the more data would only make the case stronger.
What I think we would see, in starting the analysis back closer to 1917 (100 years being a nice round number), is that, when including a binary factor like "organized crime related," that this factor would explain a lot of the variance back during Prohibition, due to the rum running criminal syndicates and cartels. A concerted effort to stop these people (via e.g. 'the untouchables') reduced mass murders, but also we repealed Prohibition, and made the National Firearms Act (NFA) around that same period, and again, each of these factors could easily be included in an analysis offering up more degrees of freedom.
I'd be very interested to see how the NFA affected murder rate. Along with this factor, we could include whether the murder in question involved an NFA weapon, to add more insight into the effectiveness of the NFA in keeping us safer, if any.
I concede you appear to have a real grasp of the topic and one that, so far as I can tell at present, is more broadly developed than my own. It's fun to read. Doesn't alter your concession or my points, but if you want to keep doing this I'm game to keep reading it in relation to the additional consideration you indicate you want to give the larger issue.
That was a very creative bit of writing, but no. It's on par with suggesting that were we to look over time at highway fatalities involving cars where the occupants were not required to wear seat belts, looked at that same data during a prolonged period where the law required the use of seat belts (we could factor non-compliance) and later looked at data after those laws were overturned and saw a lowering of fatalities during the time of the laws and a rise in fatalities once the laws were repealed, that something about seat belts led to the surge of fatalities, instead of the independent variable, the absence of those laws.
To be a good parallel, we would have to further imagine that all along, independent of the seat belt law, the number of driving fatalities was steadily rising, and that this trend overwhelms any effect due to the seat belt law, which by comparison explains much less than the general upward trend itself does. As far as I can tell from the abstract of the study in question, there is a statistically significant upward trend in the proportion of murders due to mass murder, that the "AWB" hardly influenced at all. It means there's a much more powerful factor operating than the "AWB" /seat belt law.
Also in keeping with this being a better parallel, the penalty must be similar. Not wearing a seat belt is a ticketable offense, a civil violation not a criminal one. An "AWB" would involve prison terms.
Also for lack of a better place to mention it, even if we decide that this study's result is actionable, we are only justified in reinstating the exact same law, since it is invalid statistical reasoning to use a model to support making a change that is not measured by the model. iow, what you want to do, is take this study's conclusion, and use it to justify a different "AWB" than the study's model actually measured. That is wholly invalid. It's like seeing that an apple a day keeps the doctor away, and then using that to make a law forcing us to eat oranges, thinking, "Close enough." Not close enough. Not statistically.
And in saying it step from the objective to the subjective, and I'd argue that you wouldn't find it insignificant were those still walking about members of your family or others you know and value.
No. That's incorrect. If my kin were murdered by a suicidal mass murderer using an array of assault weapons, I still would not support gun control. Suicidal mass murderers (and the other varieties also) harden my conviction that gun control is essentially unjust.
It is subjective, I grant you that, but my opinion would be if possible even more in support of us just obeying the Second Amendment, if my own kin were victims.
That is, in hybridizing your response you accidentally shed an essential humanity that is at the foundation of caring about the point at all. And in that sphere all human life is significant. The avoidance of dozens of deaths and hundreds of injuries in a given year is a compelling public good.
Ceteris paribus, I agree with you. But all other things are not equal, namely the difference between what you mean by The Right to Bear Arms, and what I mean by The Right to Bear Arms. And while I'm sure that you also found your view in the appreciation for innocent lives being preserved, I do too. It is never just for an innocent person to be outgunned by their attempted murderer. Never. It is flagrantly infringing, not their right to bear arms, but their right to live, to make laws that increase the chance that innocent people will be outgunned by their attempted murderer.
No, it's more an addition to an easier and straight forward notice of rates of mass shootings, violence and homicides and how in those states and nations with the strongest gun laws the lowest levels of those undesirable elements occurs, predictably.
As mentioned, I value minimizing murders above all, and wrt murders, what you say here is not true. You continue to resist the fact that wrt murders, our great state of New Hampshire is among the safest places in the whole world, including those Western democratic cousins you allude to, and NH also has perhaps the least gun control in the whole world too. And it's also surrounded by states with similarly weak gun laws.
But yes, it is an argument. There's just a better, stronger one to be made and for stronger action.
I disagree with you, and you don't have the statistics to back up your view either. It's just a view.
A full examination of the study seems reasonable at this point to get at a number of points/concerns you raise that seem completely valid to me, though none of them seem set to nullify the conclusion or your acknowledgement of it, if indirectly.
Let me insert here something that's worth mentioning at some point, and that's comparing mass murders according to their legal definition, which includes attempted murder as equal to murder.
Wikipedia's probably good enough for the point.
Braddock killed 58 people but there were also 422 people injured from gunfire, so just prima facie that's actually 480 murders (422 attempted) he committed. By comparison, the next nine most deadly mass shootings in the US were 102 for the Pulse shooting, 49 at Virginia Tech, 28 at Sandy Hook, 46 in Sutherland Springs, 50 at "Luby's" (1991), 40 at "San Ysidro" (1984), 48 at UT (1966), 34 at Stoneham Douglas, and 46 at Fort Hood. Those nine added together were 443 actual (fatalities) and attempted murders, so it's right to see that horrific crime of Braddock as something categorically different from everything else. It was roughly ten times worse than everything except for the Pulse, it was 'only' 4-5 times as bad as that.
I'll see how easy it is to snag the thing in full and if I can I'll throw a copy your way.
If you do, I will examine it.
Not bad. I'm enjoying it, largely for your approach to it.
A safer citizenry would entail those less imperiled by gun violence and the corresponding rates of violence. And when we look at the data across generations now, every society with stronger gun laws does a better job of making their citizenry safer.
No, that's completely false. Brazil, Honduras, Guatemala, Jamaica, and the list isn't even close to complete there, are way more dangerous than the US. Many of the migrants knocking on our southern border's door are fleeing just those countries, all of them with far stricter gun control than the US has---as I mentioned previously, if you just take civilian gun ownership as a metric for the inverse of strong gun laws, no other country on the earth even approaches the US, and yet there are over 100 countries with higher murder rates, and some of the worst ones are around ten times the US murder rate, even with far less than 10% as many civilian owned guns as in the US.
In addition, I've calculated that if black men in the US would only murder proportionally to their proportion of the population, the murder rate in the US would drop by 30%, making our country even safer, even with the gobs of civilian owned guns floating around.
The stronger the laws the better the outcome, but even the weakest of our cousins in that regard do a great deal better than we manage, just as our states (if to a lesser extent) with stronger gun laws our perform states with the weakest.
I found five whole states with 'weak' gun laws and excellent (low) murder rates.
This isn't about the number of guns, though you can make an argument on the point. It's about the kinds of guns, aids, and public responsibility.
I know that. But it's not Not about the number of guns either. According to your view there should never be countries where civilian gun ownership of any kind is extremely low, and murder rate extremely high, and yet there are numerous examples of this very thing. It would be one thing if there were just one, or a small handful of countries that defied your supposed 'rule,' but there are numerous, which just reasonably challenges that it is a rule at all.
As mentioned, writ large, the number of guns of all types in civilian hands by country, shows a statistically insignificant inverse relation with murders. Since it's statistically insignificant, the justified conclusion is that, within the range of the analysis (all the way from zero guns on the one side, and 120 guns on the other, per 100 people), having more guns in civilian hands will not increase murder rates, and having zero guns in civilian hands will not lower it either.
My argument from the beginning is that we can retain the right, but that we have created a class of affordable weapons largely if not entirely distinguished by their ability to kill a great many people in a very small window of responsive time and that if we address those weapons, and couple that with mandatory gun safety courses and the elimination of certain aids that help those and other guns in the infliction of that sudden potential for mass killings, the ease of it, we can save a great many lives that we've seen lost in parks, churches, etc.
The data doesn't support it. The data supports that the more people with more guns, will not increase murders, and everybody having no guns, will not lower murders either.
Now, statistics aside, here is another possible way to fix our broken politics on this matter. I've already said that, we should amend the Constitution if we can, to revise what We the People mean by, that The Right to Bear Arms Shall Not Be Infringed, so that we aren't biting our nails hoping the Supreme Court agrees with our own view, and voting for presidents based purely on this one factor (which types of judges the president will nominate for future S. Ct. justices).
I propose that we amend the Constitution first, to increase the burden for confirming a S. Ct. justice instead, to a super majority of both houses, rather than a simple majority. Whenever a super majority is required for change, it tends to stabilize things. This will rule out judges that are in any way extreme from sitting on the highest court. And it might be easier to achieve the super majority required to amend the Constitution with such an effort, instead of getting into the details of the Second Amendment.
I still believe that in the long run, we are better off obeying our Bill of Rights, and if we don't like the Bill of Rights, then we should change our Bill of Rights, instead of making laws that flagrantly violate it, but since changing the Bill of Rights seems unlikely anytime in the foreseeable future (short of some massive disaster that shocks one side or the other into capitulation---this would be a very bad case that I do not wish for), I suggest raising the bar for confirming S. Ct. justices instead.
And all that aside, we do agree on at least one thing. Neither of us wants to see anymore mass murders, ceteris paribus. Neither of us ever wants to see again something like Braddock. Your solution to Braddock and mine are certainly wildly different, but we do both want the same outcome here. Whereas your attempt is to ban the weaponry that he used, mine is about preserving the right of the rest of us to arm ourselves at least similarly to (if not better than) how he armed himself. Mine also requires that We the People take more individual responsibility to take care of ourselves, and our kin, and our neighbors and other innocent (of capital crimes) people.
It would certainly be nice if we only dealt with reasonable people, but the reality is that sometimes people just 'lose it,' whether it's explainable through a physiological problem or not, and your attempt involves trying to prevent these people from being armed too well to begin with, and mine involves civilians being free to be better armed instead, to deter yes, but primarily to actively defend against a mass murderer (suicidal or otherwise).
In the case of Braddock, I would have preferred to see maybe dozens of innocent people in that crowd armed with rifles that could shoot back at his perch. Would it have been more chaotic, with bullets zipping back and forth instead all just in one direction, sure. Would it have involved innocent people in that hotel getting hit and killed by accident, sure. But would it have resulted in fewer fatalities? We can't know. But my guess is yes. 58 dead, and 422 wounded from bullets (480 casualties not including those injured by other means), is what we had, with nobody in that crowd shooting back. I think it would have been better had some people in the crowd been able to return fire. And they simply could not have done that with handguns, but only with rifles, due to the range involved. Handguns can shoot bullets that far, but only by 'lobbing' them (ballistically), and by the time they would have got to Braddock (in the unlikely event that they actually got to exactly his room) they wouldn't even be going fast enough to disable him.