John De Armond said: Kennesaw is the city. Even though the law is symbolic, it served its purpose. Burglaries dropped to zero the following year. That's ZERO. Nadda. Gee, this story gets better every time it is told. Next time it is repeated I suppose we will hear about how the the Kennesaw gun law caused the rate to become negative. Here are the actual numbers (from Sociology & Social Research v74:1 p51) Kennesaw Burglaries 1976-1986 76 77 78 79 80 81 82 83 48 85 86 41 21 22 35 35 54 35 35 29 32 70 The Kennesaw law was passed on March 15, 1982 and pretty clearly had no effect on the…

Your claim that I have not shown that the situations were stable is false. The homicide rate was roughly constant in the period before gun control and in the period after gun control. Andy Freeman said: The graphs have shown that it was roughly constant AFTER, but before.... there was a dip in the period 1915-18, associated with WWI and fluctuations before 1905. It might be that demographic change caused the decline i.e. a decrease in the percentage of young men in the population. However, the demographic change associated with WWI when 40% of the males 18-45 enlisted, is far far larger…

Alan Watt said: However, what effect did WW-I have on the age-distribution of the population? I would expect the percentage of 18-25 year-olds in the general population to be reduced due to war casualties. In the U.S., this is the age group which accounts for most of the violent crime. Here are the percentages of the population of NSW that were between 18 and 25 inclusive at each census year: Male Female 1911 8.4% 8.2% 1921 6.4% 7.1% 1933 7.3% 7.0% This suggests that the male percentage would have been 7.4 in 1921 were it not for the Great War. However, I…

What on earth do you mean by 'the "nothing else happened" parameter"? Andy Freeman said: Lambert's model is for a transition between two stable situations with some "noise". He uses it to argue that gun control explains the transition. Yet, he doesn't bother to show whether or not anything else happened at the relevant time, whether or not the situations were in fact stable, and so on. You call this a "parameter"?? This is a bizarre usage even by your standards. Your claim that I have not shown that the situations were stable is false. The homicide rate was roughly constant in the…

Crime rates go up and crime rates go down. Before seizing on some possibly coincidental factor such as gun training or gun control as the cause of the change, we need to establish if the change was unusual, i.e. statistically significant. The only attempt I have seen to establish this is in Kleck and Bordua's paper which claims that the change was significant since it exceeded two standard deviations. This is wrong. A rate two standard deviations from the mean would be significant, but changes exceeding two standard deviations occur 15% of the time for normal variates, nowhere near the 5%…

robert i kesten said: Could you give the date(s) and an 11 year (5 before, first year of implementation, 5 after) table of homicide rates. If possible I'm interested in Australia as a whole, not just NSW or any other state. Here are the states for which I have data. (Data for WA and Tas are incomplete in ways that make them useless for evaluating gun laws.) Qld and SA figures come from "Source Book of Australian Criminal & Social Statistics 1900-1980" Mukherjee. Victoria from "Victorian Year Book" volumes for 1901,1902,1903,1904, etc. NSW from "Homicide:The Social Reality" Wallace. -…

We've been around on this before, and all it does is impress me with the predilection of some pro-gun folks for self-delusion on this topic. (I'm sorry if that sounds harsh, but it seems to me that many people suspend their powers of reason on this issue.) Here are the NSW homicide rates from 1910-1920: 2.6 2.7 2.4 2.1 2.4 2.6 2.1 1.9 1.4 2.7 2.7 Care to demonstrate a decline? Geoff Miller said: What can be demonstrated very easily is that you are supporting your own position with some very carefully chosen figures. Untrue. Frank Crary claimed that there was a decade long decline, ending…

If you want to consider population density, Alaska has a density 7 times that of Yukon. This is a rather enormous difference. Andy Freeman said: But, is it a significant one? The relative size of the empty spaces probably doesn't matter much, except when it comes to computing average population density, because we really can ignore places where there's no one around to kill or be killed. I think it is up to those who claim that the two places are comparable, to show that, ignoring uninhabited areas, the densities are the same. Here is another way they differ: % of population living in…

Andy Freeman said: Since Alaska is significantly larger, that factor of 20 is not particularly relevant. If you want to consider population density, Alaska has a density 7 times that of Yukon. This is a rather enormous difference. Furthermore, gun availability may well be HIGHER in the Yukon. (Centerwall does not have any data on gun ownership in Alaska and Yukon.) Legal availability is certainly lower in the Yukon than in Alaska. It is? What evidence do you have?

Dean Payne said: Centerwall made his comparisons with and without the major (pop. > 1M) metropolitan areas. With these areas, I get the same numbers you list. Without, I get 3.1 for Canadian provinces, and 3.7 for the US states. I get the same numbers. Here are the homicide rates, inside and outside major metropolitan areas. homicide handgun % with rate homicide rate handgun Canada 2.8 0.3 11 <1M 3.1 0.2 6 1M 2.2 0.4 18 US 8.5…

(BTW: I find your claim that "the difference in the murder rates is explained by the different racial fractions in each city" rather strange, when Centerwall has shown that when household crowding is controlled for, black and white domestic homicide rates in Atlanta are the same.) C. D. Tavares said: Huh? 'Scuse me, which one is Atlanta in: Seattle, or Vancouver? The situation in Seattle may be different from Atlanta (and the black and white homicide rates in Vancouver do not differ significantly), but that is not the only piece of evidence that socio-economic status explains racial…

Andy Freeman said: the murder rate is HIGHER in comparable regions of Canada than it is in the US. See Centerwall's paper in the Dec 91 issue of the American Journal of Epidemiology. Seattle and Vancouver ARE apples and oranges. The difference in murder rates is explained by the different racial fractions in each city. Could you tell us what definition of "comparable" you used where Alaska and Yukon are "comparable" and Seattle and Vancouver are not? Read the paper; it isn't my definition of "comparable". I have read the paper. By the definition that Centerwall uses (geographical…

The Terminator said: In England, the percentage of burglaries committed when the occupants are at home is something like 30%, while in the US, it's around 9%. Let me add two more data points that I was able to find: Canada (Edmonton) 10% (Canadian Urban Victimization Survey #9) and Australia (Victoria) <10% (Burglary, a Social Reality). Obviously, no conclusions about cause and effect can be reached by looking at these stats alone, however, the desired conclusion can be reached by looking at how jailed perps responded in interviews. Fear of encountering an armed victim WAS important to…

Wayne J. Warf said: I just wonder if this was a little fishing expedition by Tim. You know, take a bunch of stats and run pairwise correlations on them and see if any pop out significant at p<.05. Of course, doing this without adjusting your significance levels skews the results tremendously, but one wonders just the same. Was there an a priori hypothesis being tested against the null or was it just "shotgun statistics". <sarcasm> Yes, I sat up late at night trying to correlate things with gun ownership. I tried the number of letters in the country's name, the number of medals won…

The Terminator said: Excluding the United States and Switzerland would make this worse. Further, do you have any justification for excluding them? Eliminating data points, simply because they don't "fit" isn't very good methodology. Because with least squares estimation, outlying values bias the results. In this case they make the correlation higher, giving a value that would be quoted by a politician, and not a statistician. I made the same mistake on an oral presentation a semester ago. I was computing a linear regression for data taken in the Millikan Oil Drop Experiment. I "threw out"…

Henry E. Schaffer said: In articles various people say things like: By the way, values of 0.48 and 0.45 are REALLY BAD. and then argue over whether these are or should be publishable, etc. In summary --- AARRGH! A correlation, in itself, is neither good/bad nor publishable/unpublishable. One needs to know the "significance level" and/or such extra information such as the design/size of the experiment/survey yielding the correlation. One also needs to know what is being claimed for the correlation (in terms of explanatory or descriptive power) as to get some insight into the reaction of a…

Rick Bressler said: The Netherlands have a homicide rate about double that of the English one, and only half as many guns.... So here we have The Netherlands at about the lowest rate of gun ownership in Europe, and the Swiss with one of the highest and the homicide rates are about equal. We really need to look at more data points.... I found ownership percentages for handguns in "Experiences of Crime across the World" van Dijk, Mayhew and Killias (1991). These are from an international victimization survey. (The survey asked about long guns too, but the book does not report the answer.)…

Wright & Rossi's survey of criminals showed that the main reason why criminals carry guns is self-defence, so a large number of the 500,000 gun assaults may be illegal self-defence uses. Rick Bressler said: I have a problem with confusing an assault with a defense. The two are mutually exclusive. At least under US law, an assault is a crime, a defense is not.... There are two meanings to the term "self-defence" in my dictionary: 1. the act of defending one's own person, reputation, etc 2. (Law) the use of reasonable force against an attacker, constituting a defence in criminal law and…

About 2/3 of the crimes where guns are used for self defence are assaults, so this is the death rate that we should use. Frank Crary said: Why? As I said, attempted murders/murders would have a much (as in, order of magnitude) higher. Even if they are only a small fraction of violent crimes, their contribution would still be as great as that from assaults. Let's see, if attempted murder/murder is 5% fatal, then there must be 500,000 of them to get the same contribution as I have assumed from assaults. That gives 25,000 homicides from attempted murder/murder and 25,000 from assaults. Oops…

The death rate from robberies is about 1.5 per thousand robberies. If the same death rate occurs in other crimes, then guns save 1.5*65, or about 100 lives per year. Frank Crary said: What makes you think that the death rate from robberies is typical? It certainly isn't for, say, murders or attempted murders. I'd expect it to be much higher for other crimes, such as rape (although there is a selection effect, about half the rapes reported by the local news are rape/murders.) By selecting robberies, and using 1.5/1000 as the death rate, your estimate is low by possibly as much as a factor of…