Maaate! Living in a country without the level of guns you have I could make the same claim about........
Critique isn't difficult. You show that the empirical analysis is not robust. Take Lott's bobbins. Easy to show that he deliberately relied on a biased dummy variable methodology. Remove the error, remove the relevance of the hypothesis test.
Arthur Kellerman is without credibility, and admitted in his flawed study that renting, living alone, and engaging in the use of illicit narcotic substances, pose a substantially higher risk of being murdered, than mere firearms ownership. Beyond that, self defense incidents that occurred outside of the home were not counted in the final tally.
No, I read the study. Haven't you? http://www.nejm.org/doi/full/10.1056/NEJM199310073291506 See Table 4.
Mea culpa However it is a well conducted study No this is part of the research findings and if he wanted to he could have ignored those factors and posted a biased study
The study is indeed biased, and deliberately constructed in a manner to present the false narrative that legally owning a firearm presents a greater risk of being a victim of violent crime.
The fact that criminals are dangerous to each other and their victims is not news. If Kellermann wanted to show that lawful citizens are at increased risk for being murdered by having a gun in the house, he should have only used lawful citizens in the data. As it stands, the study is worthless to make any observations about lawful citizens and gun ownership.
Why would you think that? The issue is how many of those counted in the increased risk pool for a gun in the house were criminals not legally allowed to iwn a gun, or how many were by a murderer who brought tgetgun in the house.
Studies that look at alll the incidents and have consistently come to the same conclusions Here is another study showing a greater risk of death if a gun is in the home https://jamanetwork.com/journals/jamainternalmedicine/article-abstract/623145?redirect=true
If the conclusion is that living in the house with a criminal who owns a gun leads to an increase homicide risk, we don't dispute that, any more than we would dispute that living in the house with a criminal or man prone to domestic violence increases the risk of homicide by any means. None of these studies look at just homes where there are just law abiding citizens who aren't alcoholics or drug abusers. Kellermann told us that living alone or renting is riskier than having a gun in the house, even if you or another resident is a criminal. Also, none of the studies quantify the actual risk. What is the true risk of a law abiding citizen and murder? Pretty dang small, which is why they only use "factors". Otherwise, the results wouldn't even matter. Do you think that households that depend upon automobiles for transportation rather than taking mass transit have a higher risk of dying in a motor vehicle accident? "Commuters Reduce Their Crash Risk by More Than 90 Percent When Taking Public Transit Instead of Driving" http://www.apta.com/mediacenter/pressreleases/2016/Pages/Hidden-Traffic-Safety-Solution.aspx
Once again, no one is able to pinpoint what is wrong with Kellermann's study. The factors that Kellermann took into consideration in his multivariate analysis: He still found an independent relationship between guns being kept in the home and an increased risk of being murdered.
Without knowing how much of the case study were criminals, it doesn't tell us anything about the risk to law-abiding.
A risk, even to criminals, much lower than living alone or renting. What is the risk to someone who is not a criminal or lives with one?
So what is a " criminal" in your eyes? Someone who has been caught littering? This is a random cross section what are the odds that a high percentage are criminals?
here is the bottom line-if you are afraid of the possible consequences of owning gun, then don't. But don't tell us who have made the decision to do so, not to.
This was a cross section of the population Are you telling me that the majority of Americans fit this profile?
I have done multivariate analysis before in R. First you start with a statistical model (an equation) like this: risk of being murdered = 20.5 + 1.92(gun ownership) + 2.56(drinks alcoholic beverages) + 3.4(gun ownership*drinks alcoholic beverages) + 2.4(gun ownership)^2 + 2.9(drinks alcoholic beverages)^2 One by one terms are eliminated from the equation if you can't be at least 95% confident that they are correlated with the dependent variable (the risk of being murdered in Kellermann's study). Of course, Kellermann did a much more complex and thorough analysis than the above example. He used 31 independent variables in his initial model. His final model only contained six independent variables linked to an increased risk of murder. Gun ownership was one of them.
Thank you for the statistical insight. I have to admit it is one of my limitations and I always enjoy when others can illuminate this like you just did Bravo,
