Deltoid

Since Scripter argues so badly, litters his postings with insults and outright lies, deliberately cooks his statistics and frequently gets his facts wrong, it seems to me that he he is trying to convince people that the pro-gun case is totally without merit. Is Charles Scripter an HCI plant?

Ingredients of a Scripter posting.

(1) Insults:

… oppressors, such as yourself, … … stupid sheep … … Lambert, much as the rapist does … … Tim “Lean back and enjoy it” Lambert.

You are but a filthy vermin that scurrys for hiding when the light is turned on. Likewise your arguments require obfuscation and darkness to thrive, and crumble when exposed to the light of scientific inquiry. As for death threats, an invertebrate such as yourself is unworthy to soil my blade…

Presumably, if he didn’t consider me an “invertebrate” he would attempt to murder me, so I guess we should be thankful for small mercies.

Scripter hopes that I will respond in kind to his insults so that the discussion will degenerate into a slanging match and we will not discuss the facts, since he knows that the facts do not support him.

(2) Lies:

Timmy has already “scientifically proven” (read: “manufactured statistical trends”) that guns cause crime,

1) You haven’t the faintest idea what underlying mathematical assumptions are made to develop the statistical formulae you use, nor do you understand the correct application of these formulae, or

2) You are deliberately and blatantly guilty of academic dishonesty.

he DOESN’T UNDERSTAND statistics

Timmy has repeatedly demonstrated his inability to solve simple probability problems.

Timmy has manufactured trends; he has ignored various potentially important parameters in his assumptions, he has failed to justify the discard of these parameters, and he has drawn incorrect conclusions from his data.

(3) Getting it wrong:

armedThe statistic is clearly marked in the study as unreliable because gun and knife self defence was so uncommon. The actual numbers were 1 completion out of 32 attempts. A 95% confidence interval for the completion rate is 1-15%. The completion rate for non-violent resistance was 10%. So the data does not tell us whether resisting with a weapon was better or worse than non-violent resistance.

But you ignore the data from a much larger study. Here’s a clip from a file I found that discusses it.

A 1979 US Justice Department study of 32,000 attempted rapes show that overall, when rape is attempted, the completion rate is 36%. But when a woman defends herself with a gun, the completion rate drops to 3%.

A sample size of 32,000 is probably much more reliable than the 32 events you have chosen to attack. Lambert, much as the rapist does, has chosen to attack a smaller and weaker opponent.

No Charles, it’s the SAME STUDY. Here is the complete reference that those numbers above came from:

AUTHOR :McDermott, M. Joan
TITLE :Rape victimization in 26 American cities
PLACE :Washington
PUBLISHER :National Criminal Justice
DATE :1977

The relevant number to estimating the reliability of the estimate for the completion rate when gun/knife defense is used is the number of samples when gun/knife defense was used, not the total sample size.

“armed woman” does not equal “resist with gun or knife”. Someone with a weapon may be unable or choose not to use it.

Non sequitur, since I specifically referred to women who used a weapon to resist.

And followed it by talking about disarming people, indicating that you thought the statistic had some relevance to the benefits or otherwise of being armed.

Bill Gray said:

[talking about Orlando gun training]

Anyway, during the period after the training when the rapes dropped so dramatically, none of the women ever fired their weapons. Moreover, none of them shot a husband or boyfriend, nor were any children harmed with these firearms.

I personally called the Orlando PD to confirm as much of this as I could, and it checks out with their records department. There was an increase in homicide the same year, but the base numbers were so low that even one or two additional would come out to the 22%–in short, it wasn’t like they dropped 122 compared to the previous 100, but more like 17 instead of the previous 14. It was not considered statistically significant by the statistician I spoke to at the Orlando PD anyway.

On the basis of what is known, it is hard to see a down side of the Orlando experience. It is pretty clear–as clear as these things ever get–that the well-publicized ability of a small percentage of the potential victims to repel assault provided protection to nearly all of that class of potential victims. It is also clear that there was no societal cost–aside from the trivial expense to the police of providing the training (which may have been offset by fees charged–I do not know if the classes were gratis or not); certainly the increase in homicide was not related to the women being armed.

To summarize your argument:

1. the increase in homicides was not statistically significant — the actual number of homicides was small enough that increases of 20% or more were not uncommon.

2. None of the women killed anyone.

Hence you conclude the increase in homicides was unrelated to the training. Fair enough. I agree with you.

However,

1. the decrease in reported rapes was not statistically significant – the rape rate fluctuated wildly in the years before the training program. It even decreased 100% one year.

2. None of the women used her gun to defend against rape.

So why are you so sure that the gun training caused the reduction in the rape rate?

Certainly, it is possible that some criminals were so scared by the news of armed victims that they gave up rape for a year or more, so the reduction in rapes, though not statistically significant might have been caused by the training.

It is also possible that some unarmed criminals were so scared by the news of armed victims that they armed themselves and ended up shooting someone, so the increase in homicides, though not statistically significant might have been caused by the training.

Jim De Arras said:

60% of all house burglaries in GB occur while the house is occupied. Less than 10% here. More than 50% of homes here have firearms, less than 5% there.

I’ve proven that the mere existance without use of most guns in homes in the USA makes even your gunless home safer to sleep in than homes in GB.

Let’s see: Your proof goes:

(1) Gun ownership is greater in the US than in GB.
(2) “At-home” burglaries are more common in GB than in the US.
Therefore:
(3) Gun ownership deters “at-home” burglary.

We also have:
(4) Gun ownership is greater in the US than in Canada.
(5) “At-home” burglaries are more common in the US (14%, NCS data) than in Canada (10%, Canadian Urban Victimization Survey #9)
Therefore (by your reasoning)
(6) Gun ownership encourages “at-home” burglary.

I don’t see how (3) and (6) can both be true. Perhaps there is something wrong with your proof technique?

Jim De Arras said:

Well, Mr. Lambert, lets have the numbers for .au, and see where the trend leads us.

Here are all the numbers I have.

Country     % at-home   % gun      homicide
            burglaries  ownership  rate
Netherlands 48           2         0.9
England     26-59        5         0.7
Australia   10          20         2.0
Canada      10          31         2.1
USA         14          49         8.8

The Australian “at-home” burglary rate is actually for Victoria. The range given for England is because the rate is 59% for attempted burglaries and 26% for completed burglaries, so the overall rate must be somewhere in between.

The GB example at least cancels the Canadian one, so (6) is disproved if (3) also is.

My point is that is nonsense to take a correlation with two data points and call it a proof.

And while not having the raw numbers, I still suspect Canada is far from unarmed, much farther than GB, and so the 4% difference between the USA and Canada is in the noise. There would be a threshold above which the assumption is that the home might be armed, and so better to wait until they are not there.

Well, if your threshold theory is true, the threshold must be somewhere between 5% and 20%. However, the same reasoning will also prove that greater gun ownership causes more homicides. Do you concede this, or do you think some other factor is involved?