There is a large supply of junk science generated by the anti-gun types. The statistics which supposedly show that houses which have guns in them are more likely to have a crime happen where the gun is used to commit the crime as opposed to stopping the crime rely on a population of criminals who, needless to say, keep their guns at home.
The studies I have read tout the correlation of gun ownership with violence in households, but ignore a stronger correlation from their own data between renting and violence in households. By the logic of the studies you could dramatically increase your safety by buying a house instead of renting one. But that defies logic, suggesting the underlying study methodology is incorrect.
One study I read claimed a correlation between gun ownership rates and suicides, but neglected to point out that they generated their estimates of gun ownership rates from the firearm suicide rate. That kind of circular statistical manipulation is not a valid scientific method.
You can buy a house in Detroit for $1 - then you'd be real safe.