Glenn Reynolds points to a page that purports to show the effect on crime of Kennesaw’s ordinance that made gun ownership mandatory. Unfortunately the numbers given there are misleading—they just give the crime rates for the year before the ordinance (54 burglaries) and the year the ordinance was passed (35 burglaries) and for 1998 (36 burglaries). This makes it look like there was a decrease, but if you look at burglary rates over a ten year period (see the graph to the left ) there is a lot of fluctuation. A statistical test shows that there was a statistically insignificant increase .
It’s not surprising that the ordinance had no effect. It was purely symbolic and was never enforced so is unlikely to have had any effect on gun ownership in Kennesaw.
There was also a large increase in the population of Kennesaw, which meant that by 1998, although the number of burglaries had not changed, the burglary rate per 100,000 population had decreased greatly. It is hard to attribute this to the ordinance since the large increase meant that the people living in Kennesaw in 1998 were almost completely different from those living there in 1981.
 David McDowall, Brian Wiersema and Colin Loftin. “Did Mandatory firearm ownership in Kennesaw really prevent burglaries?” Sociology and Social Research 74:48-51 (1989).
Update: Glenn Reynolds responds in an update:
I don’t think that burglars check resumes, and I don’t see why duration of residency should make any difference at all here. And if the number of burglaries stays the same, while the population grows, that means that burglary is getting less common. Doesn’t it?
The burglary rate decreased gradually as the population increased. It doesn’t make sense to attribute a gradual decrease to a one time event like the ordinance. If it had an effect, you would expect an abrupt decrease when it was passed, but there was no such decrease.