>>>>> On Tue, 1 Dec 1992 21:02:15 GMT, oleary@cbnewsh.cb.att.com (brian.m.leary) said:
> (Tim Lambert) writes:
>> See the article in Criminology v29:4 (1991) pp 541-559.
>>
>> The raw data is presented as graphs showing burglaries in Morton Grove
>> and Kennesaw for the period 1976-86. The burglary rate in Kennesaw
>> appears stable, while the burglary rate in Morton Grove is noticeably
>> lower. The paper's authors did an interrupted time series analysis on
>> the data: the change associated with the Kennesaw ordinance was an
>> insignificant increase of 0.2 burglaries/month, the
>> change associated with the Morton Grove ordinance was a
>> statistically significant decrease of 4.5 burglaries/month,
> Please note that the burglary rate is in burglaries/month.
> Other adjustments might be needed. For example, in TPG number of
> crimes divided by population over some unit time is a more common rate.
>> Clearly the Kennesaw ordinance had no effect. It seems unlikely that
>> the Morton Grove ordinance caused the reduction (unless there a
>> significant number of burglars who burgle primarily to get guns). The
>> authors also considered the effect of the Evanston gun ban on
>> burglaries and found a not quite significant (p=.13) decrease.
>>
>> The data does not support the theory that gun ownership deters
>> burglaries.
> Please note that Mr. Lambert is not presenting the data, he is
> presenting someone's analysis. We don't know what the data supports.
The data is presented as a graph. You don't need a whole lot of
analysis to see that there was no decline in Kennesaw.
> I would suggest some re-analysis of the data. I don't have the report
> so perhaps Mr. Lambert could help.
I can scan the graphs and post GIFs if that would be any help. I've
already posted annual totals for Kennesaw. Here they are again:
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
> First, I think that making the rate population based might change the
> analysis - I think that Kennesaw grew quite a lot
> while Morton Grove was stable or shrinking. For example, if Kennesaw
> doubled in size while burglaries/month remained constant the burglary
> rate would be halved. I would hope that the authors did this but
> after seeing the DC study ....
It is not clear that rate per 100,000 pop is the correct measure.
Most burglars are young males, so rate per 100,000 young males might
be better. Maybe burglars come from outside Kennesaw. Then rate per
100,000 residences (number of places to burgle) is better. Burglars
tend to be poorer. Perhaps a whole lot of wealthier people moved to
Kennesaw.....
Fortunately we can distinguish the effect of these factors from the
possible effect of the gun law. The gun law is supposed to have had
an abrupt effect, while the other factors would operate gradually over
the whole period. If some other factor (such as increasing
population) caused the number of burglaries to double over the ten
years, canceling out a 50% drop caused by the law, then the graph
would look like this:
*****
***** *
****
**** * *****
****
*****
It doesn't. Interrupted time series analysis is designed to detect
just this kind of thing and found no reduction associated with the
law.
If an increase in population masked the effect of a 73% reduction in
rate per 100,000 pop then the population of Kennesaw must have
quadrupled in the one year following the gun law. I think this
unlikely.
Feel free to look up the population figures and repeat the analysis.
The result will be the same.
> Second, I think that the residential burglary rate should be used;
> not the total burglary rate. I don't think that a law requiring
> residents to own guns would have much impact on the burglary
> of businesses. Which rate did the authors use?
I don't have the paper with me at present. Even if the figures are
for total burglaries, residential burglaries make up the major
component of this, so if residential burglaries fell 73%, total
burglaries should also fall significantly.
> Third, did the authors of the study cited compare the stated decline in
> Morton Grove with State and National trends - I seem to remember the
> crime rate in the US declining in the early eighties. Did they
> adjust the data for population changes? How about
> Kennesaw - what happened in the surrounding communities?
The Kennesaw and Morton Grove data cover the same time frame. If the
national trend was downwards, Kennesaw did not follow it. The authors
did not do comparisons with surrounding areas. They were after all,
demonstrating that there was no effect, not considering other causes
for an observed effect.
> Making these adjustments will probably result in the large
> - population adjusted residential - burglary rate decreases
> published in various gun magazines - though probably not 400 percent.
I very much doubt it. I tend to believe articles published in
refereed criminological journals that provide the data, explain the
methodology, and whose analysis I can check more than I believe second
hand reports of stories in magazines that report impossible results.
Tim