The family that posts together

In response to criticism that Lott used cherry picked numbers to claim that homicides had increased following gun registration in Canada, "Maxim" posted on Usenet:

The law started in December 1998. Guns did not have to be registered until 2001. Violent crime was falling until very shortly after the law started doing anything, and the crime rates start rising consistently after that, with the biggest increase in 2002. It is falling before the law and for one year after and then rising consistently afterward.

Hmm, it was posted under the name of Lott's son, but the writing style belongs to John Lott. Note that Lott has posted a five-star review of More Guns, Less Crime under Maxim's name. I suppose it's possible that John helped Maxim with the post.

John Lott posted on his blog:

Canadian long gun and shot gun registration was started in December 1998 and the guns were supposed to be registered by January 2001, though there was an extension until July 2001. Crime rates continued falling during the first full year of the program but rose consistent [sic] every year thereafter and in the lastest [sic] numbers for 2002 ended up higher than the last year prior to the law (1998).

Maxim, at least, seems to be better at proof reading his posts.

i-dec5473c4486bf930415768466401c43-canhom.png Anyway, on to the substance of their defence. The "consistent" increase is barely noticeable on the graph since it is an increase from 1.76 to 1.77 to 1.78. Given the normal random variation in crime rates, no person knowledgable in statistics would read anything into such an increase.

Furthermore, when you look at the official Canadian statistics you'll see that there is a footnote on the 2002 homicide rates:

Part of the increase in 2002 is a result of 15 homicides that occurred in Port Coquitlam in previous years and that were reported by police in 2002. Homicide counts reflect the year in which police file the report.

If you adjust the the statistics to count those murders when they happened, the rates for 2001 and 2002 end up exactly the same at 1.80 each, and even Lott's cherrypicked increase goes away.

More like this

Lott has teamed up with Eli Lehrer for another cherry picking exercise. In an op-ed drafted in National Post they get straight to it with an outrageous cherry pick in just the second sentence: Gun control has not worked in Canada. Since the new gun registration program started in 1998…
Last year blogger Xrlq href="http://xrlq.com/2004/03/17/hat-of-the-day-stim-lamberts-xrlq/" rel="nofollow">dismissed my criticism of Lott as "paranoid rantings" and "gratuitous attacks on Lott personally", calling me "Dim", "Timwit", "Timbecile", "a jerk" and "Dim Lambert". This year I noted…
This is an annotated list of John Lott's on line reviews at Amazon and at Barnes and Noble. Most of his reviews were posted anonymously or under a false name, and he used this anonymity to post many five-star reviews of his own books and to pan rival books. When you post a review at…
Lehrer and Lott have recycled their previous cherrypicking exercise into an article in the Investor's Business Daily falsely claiming that gun control in Britain, Canada and Australia have lead to "historic increases in crime". Mostly they repeat their previous claims, so I'll just…

Statistics Canada reported yesterday:

"The national homicide rate fell 7% last year to its lowest level in over 35 years. A total of 548 homicides were reported to police, 34 fewer than in 2002."

By Carl Jarrett (not verified) on 29 Jul 2004 #permalink