Sam A. Kersh writes:
No, it up to you to show that an 88% reduction in rapes, 35% reduction in commercial robbery and an 84% reduction in residential burglaries is insignificant. Proof, not an inane assertion by Lambert as a citation.
Sam, you seem to lack the most basic understanding of the way statistics is done. Changes are not assumed to be statistically significant until proved otherwise. The burden of proof is on you: present a statistical test for significance. I’ve asked you for this several times now. You keep trying to avoid the issue. (Hint for Sam: posting extracts from Kleck which contain no such test is not presenting a test for statistical significance. Clue for Sam: Kleck knows how to tests for statistical significance. Why doesn’t he present one for this case?).
Tim, you made an unsupported statement that a change of 88% in number of rapes was not statistically significant.
Unsupported? I have provided plenty of support for my claim. See: here. I have also pointed out that McDowall et al (Criminology 29:4 p541-559) did an interrupted time series analysis. They did find a reduction by 11 rapes in 67, but this was NOT statistically significant.
I’m still waiting for you to provide any sort of support at all for your claim that the reduction in Orlando was statistically significant
Heck, you can’t even support your claim about robberies in KC. Why don’t you tell us how many robberies there were before and how many there were after?
Funny thing, neither of you developed data supporting your assertions.
Totally false. We have provided plenty of support for pour assertions. When your bogus claims are taken apart in detail you merely repeat them and pretend the refutation does not exist.
Yes, I’ve repeatedly pointed you to Kleck and Green (1987, p75)
Oh, Green. Let’s see what he says:
“Regarding the authors’ [Kleck and Bordua] first contention (that the reduction is too great to be considered random), the City of Orlando seems to have experienced a rather jagged yearly rate history for recorded forcible rape (including attempts) in the years prior to the firearms campaign. For example, one infers from any recorded rape/attempted rape of 0.0 (in 1963) for a city as large as Orlando that the recording procedures there are questionable. The recorded decrease in the City of Orlando from 1959-1960 was 58%, the decrease from 1961 to 1962 was 88%, and of course, the decrease from 1962-1963 was 100%. Recorded fluctuations in rape in Orlando from 1964-1966 are similarly extreme. Hence, the authors’ assertion (1983: 287) — that Orlando had not experienced a similarly large decrease prior to 1967 than in 1967 — is misleading, given the changes in the rate recorded in, say, 1962-1964.” (Criminology 25 p 74)
Gee Sam, did you really think you could get away with attempting to pass off Green as supporting Kleck?
but you two keep going back to Mcdowall’s flawed studies that have no clear relevance to Kleck’s 1988 study.
McDowall tested the statistical significance of the change and found that it wasn’t statistically significant. I hope that makes the relevance clear to you now.
Yes, I know McDowall as late as 1991 was trying to disprove relevance for changes in Orlando rapes and Kennesaw burglaries but he, like Kellermann, keeps getting caught using incorrect data to foster his predetermined assumptions.
McDowall used UCR data for Orlando, just like Kleck. The UCR data may indeed be incorrect as Green argues. But then your whole case, which is based on that UCR data, falls apart. As for Kennesaw, I have already demonstrated the falsity of your claims. See [here](http://timlambert.org/category/guns/kennesaw/
But what can I expect from someone who falsely tries to attribute a “200,000 killed or wounded” in DGUs computation to Kleck.
I did no such thing. What I said was that if all of Kleck’s respondents told the truth then there were 200,000 DGU woundings.