And yet, Kable remains. Is he really too stupid to figure out he’s in way over his head?
This would be funny, but it’s so sickeningly wrong. You see no evidence of this? So, you are calling bullshit on the O.P. and the entire store? If so, then I can understand why you see no evidence of this.
However, I and many other posters to this thread do believe this to be a true situation. If it is true, then the fucknugget who was dumb enough to handle and clean a weapon in front of his child is clearly possessed of a misplaced set of priorities.
Got it? Good.
Yes. He is the living embodiment of the Dunning–Kruger effect.
I think what researchers often say is something to the effect of “we were unable to find statistical significance, however there was a trend in the direction of…”
However if there the trend is in a direction they don’t like, then they don’t say things like the above quite so much.
But in his study participants he did. He just couldn’t say with arbitrarily enough certainty if that association is indicative of the population at large. Except that with rifles and shotguns he did still implicate them the other direction in his conclusion. What a dishonest dickhead Kellerman was, and you are.
Damuri, is obviously the only smart one around here. 
Lets you out.
No, what i meant was, I see no evidence that he loved his guns more than he loved his child. Its a horrible things to say. The only excuse for saying something like that is if you are possessed by such an irrational fear of guns that you think that the mere handling of firearms in the presence of your children means you don’t really mind them getting shot through the head.
Once again, putting aside the mutlivariate stuff, it seems to me that the word “significant” might have a special meaning in statistics that is different from the common understanding.
Insults seem to be a pretty important part of most of your arguments.
If this is news to you, pal, then we really do need to take a giant step back. Especially since this has already been discussed.
I would wager that the ratio of didactics to insults in my gun posts has been far, far greater than anyone else’s on this topic, and probably above the average on most all topics discussed on the boards. I know it sure feels that way to me.
When I do employ insults, it is the result of frustration. For instance, it’s frustrating after all this to have you suggest that maybe there’s a specific meaning to “significant” in a statistics context.
OK so I went back to where you discussed it and by definition significant does not include any results that include the zero effect in the confidence interval.
It still doesn’t mean there isn’t an effect. The fact that the zero effect is off to one side of the confidence interval still means something and is “significant” in the colloquial sense, right? Because the way you present it, I was getting the impression that the fact that the confidence interval barely includes the zero effect meant nothing. I thought this was the main point that Kable was trying to make.
You may feel that way but there are people who haven’t made a personal insult during the entire debate.
Its still not a very convincing tactic.
Go back to where I explained that the decision to determine whether or not an effect is significant at a p-value of .05 means that we are willing to accept that if we had collected 100 samples, we would see the magnitude of the effect we are looking at (or greater) in fewer than 5 of those samples just by random chance.
No. In fact, it’s probably the opposite of significant in the colloquial sense.
I don’t think there’s any way to say it any simpler than we can be 95% confident that the effect here is zero. I don’t know how many times I’ve said that, but how are you not getting that? Doesn’t that in the colloquial sense compute for you?
Let’s look at the actual values in the Kellermann study: 12.2% of those with a homicide in the home had a rifle in the home; 13.9% of those in the control group had a rifle in the home. In this sample of ~388 people or so, is that a meaningful difference? 1.7%? Eyeballing it (from the colloquial perspective), I would say not really. It seems like if you drew another sample, just a handful of people going the other way would change the results.
But the point of inferential statistics is to be more empirical than just “eyeballing.” We can look at the normal distributions to know how often certain values should show up, and we can figure out the probability that a difference as big as this would show up. If we would expect it to show up by chance fewer than 5% of the time, we’re willing to say that it is significant.
In the case of deciding whether or not we reject the null hypothesis (that there is no meaningful relationship between rifles and homicides in the home), it pretty much means that yes, it does not matter where the zero effect value falls. If it falls anywhere in the confidence interval of our statistic at all, we cannot claim otherwise.
So if we then used a p value of .1 and the zero effect fell outside the confidence interval then would we be able to say that there is a 5-10% chance that there is no significant effect?
No it doesn’t. 
It sounds like youa re saying that as long as the confidence interval includes the zero effect then the entire confidence interval is zero effect.
I understand that if the zero effect falls within the confidence interval then there is a chance that there is in fact no effect. But doesn’t the fact that the zero effect is off to one side of the confidence interval say anything about the likelihood that there is in fact no effect?
That makes sense intuitively…but not mathematically. It “feels” like it ought to be true. But as noted above, math doesn’t allow “eyeballing.”
What you’re looking at is a kind of meta-statistics, where one could (entirely validly) explore the distribution of data within the confidence interval. This, however, would require taking a new poll of the data, because, as presently given, we don’t know that. The data might all be clumped tightly against one end of the interval…or clumped tightly against the other end.
Mathematics doesn’t deal in “it might be.”
1 + 1 = 3, if you sort of squint a little, and try really hard to make it so.
**
Damuri Ajashi**, I’m afraid you’re using neo-con logic here - “if the facts, science and logic don’t fit your thesis, then just claim that “common sense” says you are right.”
Bullshit Hentor, and for some reason I think you know it.
-Shrug- I’m not the one who murdered his kid. If nobody has murdered their kid, you could call my statement irrational.
As it is, as we are all firmly in agreement of, it is the absolute truth that this man murdered his son.
Before you go calling people’s fear of gun owners irrational, tell me something. Before the bullet entered that child’s brain, do you think he had a moment of irrational fear? And would you sit there criticizing him if he did?
Irrational? Please. :dubious:
You’re losing it, bub. As has been pointed out, if there were actual math errors, this would be pointed out. The mathematicians and statisticians in the room are not on your side. You can snort “bullshit” all you want, but until you show us the math, you’re just a child whacking on a lamppost with a stick.
I took stats 270 in college. I know enough to know that Hentor is full of crap. It’s not my fault if you aren’t smart enough to notice. He should have said there is a 95% chance the real value was within the interval, not 95 percent the effect was zero. The latter, like I said, is bullshit.
Nobody can repeat Kellerman’s math because he never released his full data set.
Actually, I think this is more correct…
http://onlinestatbook.com/2/estimation/confidence.html
…but it’s wordier.
Yahoo! Answers. Interesting. While you’re at it, I wonder if you could also inform as to how babby is formed?