Charles Scripter wrote:
[…regarding Tim Lambert’s assertion of a significant decrease in homicide in New South Wales in 1920, coincident with the enactment of a gun control law in NSW in that same year…]
The Lambert analysis method clearly shows that there was a significant decrease for any year one may choose, from 1910 to 1930.
The statistical test shows that we can reject the null hypothesis of no change in the homicide rate in NSW between 1910 and 1930. Because the change is so highly significant, it also possible to reject the same null hypothesis by putting a step at other years near 1920. Only Charles Scripter could look at 20 statistical tests, all showing a significant reduction in homicide, and conclude that there was no reduction.
Chris BeHanna wrote:
Oh, come on, Tim! Talk about lying by telling half-truths! You have repeatedly inferred in this forum that the “significant reduction” was caused by the gun control law
I have never claimed that the reduction was definitely caused by the gun law. Never. However, no other plausible explanation has been offered. If you have one, please share it with us.
(or else why would you call out emphasis to the “significant reduction” in 1920, when it can be clearly demonstrated that your method shows a “significant reduction” for any year from 1910 through 1930?).
I’m sorry if you feel that comparing the situation before the gun law with that after is somehow invalid. Perhaps you would share with us the Behanna-approved methodology for evaluating the effects of a law? You also persist in your misrepresentation of my “method”. A test for the significance of a reduction is not the same as a method for estimating when the reduction occured. Despite being corrected on this, you repeat the falsehood.
Charles’s point, AND YOU KNOW IT, is simply that there is no possible way that your statistical method can show or infer support for the hypothesis that the NSW gun control law of 1920 was responsible for the “significant reduction” of homicide during that same year.
Charles has repeatedly maintained that there was no significant reduction in homicides in NSW. He has “supported” this contention by a succession of bogus and erroneous arguments. His latest argument seems to be that because he did 20 tests all showing a significant reduction, that this somehow proves that there was no reduction.
It does you no credit to stick your head in the sand alongside his.
Now, having been caught in your half-truth/whole lie AGAIN, you backpedal to this “null hypothesis” nonsense.
Do as all a favour: Go to the library, get a book on elementary statistics, look up “null hypothesis”, learn what it is, and come back when you have a clue.