I could give a flying fuck how many guns your gf has. What have I written that would make you think I would disapprove of your gf having 300 guns, ffs? I don’t see any reason for me personally to ever handle one, and the vast majority of the women I know who have firearms in the home feel the same way. And no, nowhere near SLC. Rural Appalachia, prime “OMG, Obama’s a comin’ fer ma guns!!!Elebenty!!1111!!!” country. Our crime is mostly illegal trade in prescription painkillers; most folks don’t even lock their doors at night. Insisting that women need to learn to use guns to protect themselves is, around here, rather silly.
I completely understand everything you are saying.
.05 is what’s commonly used but is arbitrary, .1 would make you less sure it would hold up, and .01 more sure. Still it does not change the fact that in Kellerman’s sample those with long guns in their home had less homicides. Do you understand that? Apparently there were not enough people shot with long guns to reach significance to be sure about how this relates to the population at large, but in his sample those with long guns were less likely to be involved in a homicide. Kellerman found that, yes?
But the odds ratio was different from one. It was .7 and .8.
Funny, I look up studies online and I find a number of them. Sure Kellerman fucked up the funding from the CDC but research has gone on.
Last time I looked it up that only held if you counted suicides. Studies in countries where guns were banned found subsequent increases in suicides by other means, like hanging, so lives were not really saved.
Going off memory long guns are involved in crimes consicerably less often (cause they are harder to carry and conceal) and accidents (because they are harder to point at yourself) than handguns. The population of the USA is definetly against the banning of handguns so gun grabbers like to take handgun stats, mix them up, and use them to bag on rifles like the AR15 which they think they have the best chance of banning, even though they are rarely involved in either crimes or accidents.
Not if you practice safe gun handling.
Well actually I do.
If you stopped there someone might have believed you.
But you kept typing and ruined it. What’s impressive is that you are either so thick that you don’t know how foolish you’re looking, or you are so full of guile you think you can baffle everyone with bullshit.
You seem to think that when you collect a sample, you can conclude that any differences you measure in the sample can be interpreted as descriptive of the relationship.
If you did a study of a new drug for depression, and at the end of the study, the people on the drug scored a 16 on the Beck Depression Inventory, and the other people scored an 18, would you conclude that the drug was associated with lower depression scores?
If you collected a sample of 100 men and 100 women, and you measured the men as having an IQ of 98 and the women as having an IQ of 101, would you conclude that women are smarter than men?
A related request: please define “confidence interval” for me.
No. He did not. If you want to assert that he did, please quote where Kellermann and colleagues said that.
Did you ever wonder why regression analyses generate “parameter ESTIMATES”? Again, please look up and provide us all with the definition of “confidence interval.”
Please state the confidence interval provided in Table 3 in the Kellermann et al. article for shotguns and rifles.
ETA:
Then I cannot wait for you to explain what the confidence intervals in the Kellermann article are, and what they mean.
Sorry I’m not so familiar with depression. Is that why you don’t own a gun? Don’t want to shoot yourself?
Maybe, I’d want to know if the difference was significant and if not I’d want to see the same test on a larger sample to see if the difference held up. Let me ask you a question. If you collected the same sample would you argue that the men in the sample did or didn’t have an average IQ of 98 and the women of 101, and that 101 wasn’t larger than 98. If you had a P-value of .06 would you at least somewhat suspect that women were a few IQ points smarter than men?
“A range of values so defined that there is a specified probability that the value of a parameter lies within it.”
Enough dodging, answer my questions now.
Be that as it may be, outlierrn, yes, he really is that stupid.
Really? I’ve been posting favorable vids of women responsibly defending themselves with guns on this forum for some time. I think this was the first one:
Though I suppose in Crazycatwoman’s neighborhood such behavior wouldn’t be considered “lady like.”
What, you didn’t have hunter safety in phys ed in 10th grade?
Where the teacher taught you to wrap the rifle strap around your left arm to steady it, just like he learned in basic training on Parris Island in '43?
This was in the mid-70s, in North Dakota. I would have thought it were more common now.
Well, not very nice, actually. My mother did shoot herself, quite fatally, on Easter morning, 1986. Messy. (.357.)
So, yeah, that’s one of the reasons I won’t have a gun in my house… Seriously, is that a bad reason?
I think it might vary by state
Other than the killing people part. He seemed to be using the guns safely, atleast until he shot himself. IOW I dont think thesort of safety classes the nra has prodivdes inoculation against crazy.
Mrs Lincoln
Could you get Wayne LaPierre to say that on camera? PLeeeeaaase???
I already explained to you - I don’t own a gun because I’m not a pussy. I feel physically and emotionally strong; I have no personal needs that a gun would fill. I’m normal, like 2/3 of the country in that sense.
EXACTLY. Thanks for finally getting it. The key issue is whether what you’ve observed is significantly different from the zero effect.
The data is the data. The issue is using statistical methods to interpret what the data mean. And I would be careful about that process because I know that deceitful morons are always ready to pounce on scientific products to misrepresent the truth.
And if your aunt had balls, I wouldn’t be surprised. If you had a p-value of .06, your confidence interval would not include the zero effect. Dummy. The effects we are discussing from the Kellermann paper are very non-significant.
Well, we already knew that your one talent is cutting and pasting. I wish you could understand what you’re pasting.
I’m answering your questions repeatedly and thoroughly. You’re just not capable of understanding and responding meaningfully. Your ignorant bluster is impressing nobody.
But you have been rambling on like a castrated pussy for weeks now. But you did once admit you liked shooting guns when you tried it, and would like to shoot an AR15 in particular. I’m just trying to figure out why you are such a pussy about it. Not being able to pass a background check would be one reason, you can’t get one, so you don’t want others to have them. I just like to figure out why people hold the views they do.
See for me the key is whether there is an actual effect or not, rather than if the researcher was able to demonstrate significance. But that don’t change the fact that in Kellerman’s sample long guns were associated with a lesser risk of homicide, .7 and .8 in fact, and I love watching how lengths you will go to, to avoid that simple fact.
Then after Kellerman failed to show long guns caused an increased risk, with the data if anything showing the opposite, he implemented them the same as handguns in the conclusion of his study. So it was good he lost his funding. Like they says their are lies, damn lies, and statistics.
Says the guy who has been desperately trying for several pages to cherry-pick a non-significant uncontrolled univariate effect as interpretable in his preferred direction.
Desperately scared and bone-deep stupid. Truly a dangerous combination.
I’m very sorry to hear that, Trinopus.
I hope you’ll pardon a digression for a moment…
So, I think for most of us, there’s little left to talk about regarding the issue of non-significance that is evident in confidence intervals.
However, a previous part of the thick-headed argument on the specifics of this particular paper involved an assertion about univariate (sometimes referred to as bivariate) versus multivariate models, or why a variable might be significant by itself in predicting an outcome but not retained when tested against other significant predictors.
This is probably something that most people intuitively grasp, but whenever it does come up, I often use the stork-baby connection.
However, I thought of another example and I thought I would test it out with you. It’s a bit rough and may not be better than the stork-baby example, but if I can shape it up it might be something I can use for teaching.
Say that, for some reason, you are doing a study of how many people go through a particular door. You’re in a public area that is open to the entire population. You measure two variables. one is whether or not people go through the door of interest. The second variable you have chosen is height.
So, ultimately you find out that there is a significant effect for height. The average height of people who went through the door was markedly, significantly, greater than the average height of people who did not go through the door. Aha! Eureka! People are more likely to go through that door if they are taller than if they are shorter, you conclude.
The problem is “the” third variable, one that you did not measure. There is one variable that, in reality, had you measured it and included it in your multivariate model, your previously significant effect for height would be completely explained away.
What variable might this have been?
The gender of the person? If it was the door to the men’s bathroom, the height difference would likely be because men are taller than women.
Yes! You would presumably find that the height difference among men alone actually has nothing to do with which men go in the bathroom and which do not.
Technically, the statistics would probably choke because gender would probably *nearly *perfectly discriminate who went in the men’s room, but I think it works as a conceptual example.
The ultimate point being that interpreting the significance of a bivariate relationship without accounting for other potentially explanatory variables is generally fraught with great difficulties.